MyArxiv

Michael Gilbert, Yannan Nellie Wu, Joel S. Emer, Vivienne Sze

Latency and energy consumption are key metrics in the performance of deep neural network (DNN) accelerators. A significant factor contributing to latency and energy is data transfers. One method to reduce transfers or data is reusing data when multiple operations use the same data. Fused-layer accelerators reuse data across operations in different layers by retaining intermediate data in on-chip buffers, which has been shown to reduce energy consumption and latency. Moreover, the intermediate data is often tiled (i.e., broken into chunks) to reduce the on-chip buffer capacity required to reuse the data. Because on-chip buffer capacity is frequently more limited than computation units, fused-layer dataflow accelerators may also recompute certain parts of the intermediate data instead of retaining them in a buffer. Achieving efficient trade-offs between on-chip buffer capacity, off-chip transfers, and recomputation requires systematic exploration of the fused-layer dataflow design space. However, prior work only explored a subset of the design space, and more efficient designs are left unexplored. In this work, we propose (1) a more extensive design space that has more choices in terms of tiling, data retention, recomputation and, importantly, allows us to explore them in combination, (2) a taxonomy to systematically specify designs, and (3) a model, LoopTree, to evaluate the latency, energy consumption, buffer capacity requirements, and off-chip transfers of designs in this design space. We validate our model against a representative set of prior architectures, achieving a worst-case 4% error. Finally, we present case studies that show how exploring this larger space results in more efficient designs (e.g., up to a 10$\times$ buffer capacity reduction to achieve the same off-chip transfers).

comment: To be published in IEEE Transactions on Circuits and Systems for Artificial Intelligence

♻ ☆ Research Directions for Verifiable Crypto-Physically Secure TEEs

Sylvain Bellemare

A niche corner of the Web3 world is increasingly making use of hardware-based Trusted Execution Environments (TEEs) to build decentralized infrastructure. One of the motivations to use TEEs is to go beyond the current performance limitations of cryptography-based alternatives such as zero-knowledge proofs (ZKP), fully homomorphic encryption (FHE), and multi-party computation (MPC). Despite their appealing advantages, current TEEs suffer from serious limitations as they are not secure against physical attacks, and their attestation mechanism is rooted in the chip manufacturer's trust. As a result, Web3 applications have to rely on cloud infrastruture to act as trusted guardians of hardware-based TEEs and have to accept to trust chip manufacturers. This work aims at exploring how we could potentially architect and implement chips that would be secure against physical attacks and would not require putting trust in chip manufacturers. One goal of this work is to motivate the Web3 movement to acknowledge and leverage the substantial amount of relevant hardware research that already exists. In brief, a combination of: (1) physical unclonable functions (PUFs) to secure the root-of-trust; (2) masking and redundancy techniques to secure computations; (3) open source hardware and imaging techniques to verify that a chip matches its expected design; can help move towards attesting that a given TEE can be trusted without the need to trust a cloud provider and a chip manufacturer.

Computation and Language 9

♻ ☆ TuneVLSeg: Prompt Tuning Benchmark for Vision-Language Segmentation Models ACCV 2024

Rabin Adhikari, Safal Thapaliya, Manish Dhakal, Bishesh Khanal

Vision-Language Models (VLMs) have shown impressive performance in vision tasks, but adapting them to new domains often requires expensive fine-tuning. Prompt tuning techniques, including textual, visual, and multimodal prompting, offer efficient alternatives by leveraging learnable prompts. However, their application to Vision-Language Segmentation Models (VLSMs) and evaluation under significant domain shifts remain unexplored. This work presents an open-source benchmarking framework, TuneVLSeg, to integrate various unimodal and multimodal prompt tuning techniques into VLSMs, making prompt tuning usable for downstream segmentation datasets with any number of classes. TuneVLSeg includes $6$ prompt tuning strategies on various prompt depths used in $2$ VLSMs totaling of $8$ different combinations. We test various prompt tuning on $8$ diverse medical datasets, including $3$ radiology datasets (breast tumor, echocardiograph, chest X-ray pathologies) and $5$ non-radiology datasets (polyp, ulcer, skin cancer), and two natural domain segmentation datasets. Our study found that textual prompt tuning struggles under significant domain shifts, from natural-domain images to medical data. Furthermore, visual prompt tuning, with fewer hyperparameters than multimodal prompt tuning, often achieves performance competitive to multimodal approaches, making it a valuable first attempt. Our work advances the understanding and applicability of different prompt-tuning techniques for robust domain-specific segmentation. The source code is available at https://github.com/naamiinepal/tunevlseg.

comment: Accepted at ACCV 2024 (oral presentation)

♻ ☆ CasiMedicos-Arg: A Medical Question Answering Dataset Annotated with Explanatory Argumentative Structures

Ekaterina Sviridova, Anar Yeginbergen, Ainara Estarrona, Elena Cabrio, Serena Villata, Rodrigo Agerri

Explaining Artificial Intelligence (AI) decisions is a major challenge nowadays in AI, in particular when applied to sensitive scenarios like medicine and law. However, the need to explain the rationale behind decisions is a main issue also for human-based deliberation as it is important to justify \textit{why} a certain decision has been taken. Resident medical doctors for instance are required not only to provide a (possibly correct) diagnosis, but also to explain how they reached a certain conclusion. Developing new tools to aid residents to train their explanation skills is therefore a central objective of AI in education. In this paper, we follow this direction, and we present, to the best of our knowledge, the first multilingual dataset for Medical Question Answering where correct and incorrect diagnoses for a clinical case are enriched with a natural language explanation written by doctors. These explanations have been manually annotated with argument components (i.e., premise, claim) and argument relations (i.e., attack, support), resulting in the Multilingual CasiMedicos-Arg dataset which consists of 558 clinical cases in four languages (English, Spanish, French, Italian) with explanations, where we annotated 5021 claims, 2313 premises, 2431 support relations, and 1106 attack relations. We conclude by showing how competitive baselines perform over this challenging dataset for the argument mining task.

comment: 9 pages

♻ ☆ Efficient Inference for Large Language Model-based Generative Recommendation

Xinyu Lin, Chaoqun Yang, Wenjie Wang, Yongqi Li, Cunxiao Du, Fuli Feng, See-Kiong Ng, Tat-Seng Chua

Large Language Model (LLM)-based generative recommendation has achieved notable success, yet its practical deployment is costly particularly due to excessive inference latency caused by autoregressive decoding. For lossless LLM decoding acceleration, Speculative Decoding (SD) has emerged as a promising solution. However, applying SD to generative recommendation presents unique challenges due to the requirement of generating top-K items (i.e., K distinct token sequences) as a recommendation list by beam search. This leads to more stringent verification in SD, where all the top-K sequences from the target LLM must be successfully drafted by the draft model at each decoding step. To alleviate this, we consider 1) boosting top-K sequence alignment between the draft model and the target LLM, and 2) relaxing the verification strategy to reduce trivial LLM calls. To this end, we propose an alignment framework named AtSpeed, which presents the AtSpeed-S optimization objective for top-K alignment under the strict top-K verification. Moreover, we introduce a relaxed sampling verification strategy that allows high-probability non-top-K drafted sequences to be accepted, significantly reducing LLM calls. Correspondingly, we propose AtSpeed-R for top-K alignment under this relaxed sampling verification. Empirical results on two real-world datasets demonstrate that AtSpeed significantly accelerates LLM-based generative recommendation, e.g., near 2x speedup under strict top-K verification and up to 2.5 speedup under relaxed sampling verification. The codes and datasets will be released in the near future.

♻ ☆ SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Masks

Fenia Christopoulou, Ronald Cardenas, Gerasimos Lampouras, Haitham Bou-Ammar, Jun Wang

Preference Optimization (PO) has proven an effective step for aligning language models to human-desired behaviors. Current variants, following the offline Direct Preference Optimization objective, have focused on a strict setting where all tokens are contributing signals of KL divergence and rewards to the loss function. However, human preference is not affected by each word in a sequence equally but is often dependent on specific words or phrases, e.g. existence of toxic terms leads to non-preferred responses. Based on this observation, we argue that not all tokens should be weighted equally during PO and propose a flexible objective termed SparsePO, that aims to automatically learn to weight the KL divergence and reward corresponding to each token during PO training. We propose two different variants of weight-masks that can either be derived from the reference model itself or learned on the fly. Notably, our method induces sparsity in the learned masks, allowing the model to learn how to best weight reward and KL divergence contributions at the token level, learning an optimal level of mask sparsity. Extensive experiments on multiple domains, including sentiment control, dialogue, text summarization and text-to-code generation, illustrate that our approach assigns meaningful weights to tokens according to the target task, generates more responses with the desired preference and improves reasoning tasks by up to 2 percentage points compared to other token- and response-level PO methods.

comment: 20 pages, 9 figures, 5 tables. Under Review

♻ ☆ DEPT: Decoupled Embeddings for Pre-training Language Models

Alex Iacob, Lorenzo Sani, Meghdad Kurmanji, William F. Shen, Xinchi Qiu, Dongqi Cai, Yan Gao, Nicholas D. Lane

Language Model pre-training benefits from a broader data mixture to enhance performance across domains and languages. However, training on such heterogeneous text corpora is complex, requiring extensive and cost-intensive efforts. Since these data sources vary in lexical, syntactic, and semantic aspects, they cause negative interference or the "curse of multilinguality". We propose a novel pre-training framework to alleviate this curse. Our method, DEPT, decouples the embedding layers from the transformer body while simultaneously training the latter in multiple contexts. DEPT enables the model to train without being bound to a shared global vocabulary. DEPT: (1) can train robustly and effectively under significant data heterogeneity, (2) reduces the parameter count of the token embeddings by up to 80% and the communication costs by 675x for billion-scale models (3) enhances model generalization and plasticity in adapting to new languages and domains, and (4) allows training with custom optimized vocabulary per data source. We prove DEPT's potential by performing the first vocabulary-agnostic federated multilingual pre-training of a 1.3 billion-parameter model across high and low-resource languages, reducing its parameter count by 409 million.

♻ ☆ Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation

Shengjie Ma, Chengjin Xu, Xuhui Jiang, Muzhi Li, Huaren Qu, Cehao Yang, Jiaxin Mao, Jian Guo

Retrieval-augmented generation (RAG) has enhanced large language models (LLMs) by using knowledge retrieval to address knowledge gaps. However, existing RAG approaches often fail to ensure the depth and completeness of the information retrieved, which is essential for complex reasoning tasks. In this work, we present Think-on-Graph 2.0 (ToG-2), a hybrid RAG framework that iteratively retrieves information from both unstructured and structured knowledge sources in a tightly integrated manner. Specifically, ToG-2 leverages knowledge graphs (KGs) to connect documents via entities, facilitating deep and knowledge-guided context retrieval. Simultaneously, it uses documents as entity contexts to enable precise and efficient graph retrieval. ToG-2 alternates between graph retrieval and context retrieval to search for in-depth clues relevant to the question, enabling LLMs to generate accurate answers. We conduct a series of experiments to demonstrate the following advantages of ToG-2: (1) ToG-2 tightly integrates context retrieval and graph retrieval, enhancing context retrieval through the KG while enabling reliable graph retrieval based on contexts; (2) it achieves deep and faithful reasoning in LLMs through an iterative knowledge retrieval process that integrates contexts and the KG; and (3) ToG-2 is training-free and compatible with various LLMs as a plug-and-play solution. Extensive experiments show that ToG-2 achieves state-of-the-art (SOTA) performance on 6 out of 7 knowledge-intensive datasets with GPT-3.5, and can elevate the performance of smaller models (e.g., LLAMA-2-13B) to the level of GPT-3.5's direct reasoning.

♻ ☆ Representing the Under-Represented: Cultural and Core Capability Benchmarks for Developing Thai Large Language Models

Dahyun Kim, Sukyung Lee, Yungi Kim, Attapol Rutherford, Chanjun Park

The rapid advancement of large language models (LLMs) has highlighted the need for robust evaluation frameworks that assess their core capabilities, such as reasoning, knowledge, and commonsense, leading to the inception of certain widely-used benchmark suites such as the H6 benchmark. However, these benchmark suites are primarily built for the English language, and there exists a lack thereof for under-represented languages, in terms of LLM development, such as Thai. On the other hand, developing LLMs for Thai should also include enhancing the cultural understanding as well as core capabilities. To address these dual challenge in Thai LLM research, we propose two key benchmarks: Thai-H6 and Thai Cultural and Linguistic Intelligence Benchmark (ThaiCLI). Through a thorough evaluation of various LLMs with multi-lingual capabilities, we provide a comprehensive analysis of the proposed benchmarks and how they contribute to Thai LLM development. Furthermore, we will make both the datasets and evaluation code publicly available to encourage further research and development for Thai LLMs.

♻ ☆ Evaluating and Safeguarding the Adversarial Robustness of Retrieval-Based In-Context Learning

Simon Yu, Jie He, Pasquale Minervini, Jeff Z. Pan

With the emergence of large language models, such as LLaMA and OpenAI GPT-3, In-Context Learning (ICL) gained significant attention due to its effectiveness and efficiency. However, ICL is very sensitive to the choice, order, and verbaliser used to encode the demonstrations in the prompt. Retrieval-Augmented ICL methods try to address this problem by leveraging retrievers to extract semantically related examples as demonstrations. While this approach yields more accurate results, its robustness against various types of adversarial attacks, including perturbations on test samples, demonstrations, and retrieved data, remains under-explored. Our study reveals that retrieval-augmented models can enhance robustness against test sample attacks, outperforming vanilla ICL with a 4.87% reduction in Attack Success Rate (ASR); however, they exhibit overconfidence in the demonstrations, leading to a 2% increase in ASR for demonstration attacks. Adversarial training can help improve the robustness of ICL methods to adversarial attacks; however, such a training scheme can be too costly in the context of LLMs. As an alternative, we introduce an effective training-free adversarial defence method, DARD, which enriches the example pool with those attacked samples. We show that DARD yields improvements in performance and robustness, achieving a 15% reduction in ASR over the baselines. Code and data are released to encourage further research: https://github.com/simonucl/adv-retreival-icl

comment: COLM 2024, 31 pages, 6 figures

♻ ☆ FacLens: Transferable Probe for Foreseeing Non-Factuality in Large Language Models

Yanling Wang, Haoyang Li, Hao Zou, Jing Zhang, Xinlei He, Qi Li, Ke Xu

Despite advancements in large language models (LLMs), non-factual responses remain prevalent. Unlike extensive studies on post-hoc detection of such responses, this work studies non-factuality prediction (NFP), aiming to predict whether an LLM will generate a non-factual response to a question before the generation process. Previous efforts on NFP have demonstrated LLMs' awareness of their internal knowledge, but they still face challenges in efficiency and transferability. In this work, we propose a lightweight NFP model named Factuality Lens (FacLens), which effectively probes hidden representations of questions for the NFP task. Besides, we discover that hidden question representations sourced from different LLMs exhibit similar NFP patterns, which enables the transferability of FacLens across LLMs to reduce development costs. Extensive experiments highlight FacLens's superiority in both effectiveness and efficiency.

Computer Vision and Pattern Recognition 9

♻ ☆ TuneVLSeg: Prompt Tuning Benchmark for Vision-Language Segmentation Models ACCV 2024

Rabin Adhikari, Safal Thapaliya, Manish Dhakal, Bishesh Khanal

comment: Accepted at ACCV 2024 (oral presentation)

♻ ☆ 3D-free meets 3D priors: Novel View Synthesis from a Single Image with Pretrained Diffusion Guidance

Taewon Kang, Divya Kothandaraman, Dinesh Manocha, Ming C. Lin

Recent 3D novel view synthesis (NVS) methods are limited to single-object-centric scenes and struggle with complex environments. They often require extensive 3D data for training, lacking generalization beyond the training distribution. Conversely, 3D-free methods can generate text-controlled views of complex, in-the-wild scenes using a pretrained stable diffusion model without the need for a large amount of 3D-based training data, but lack camera control. In this paper, we introduce a method capable of generating camera-controlled viewpoints from a single input image, by combining the benefits of 3D-free and 3D-based approaches. Our method excels in handling complex and diverse scenes without extensive training or additional 3D and multiview data. It leverages widely available pretrained NVS models for weak guidance, integrating this knowledge into a 3D-free view synthesis approach to achieve the desired results. Experimental results demonstrate that our method outperforms existing models in both qualitative and quantitative evaluations, providing high-fidelity and consistent novel view synthesis at desired camera angles across a wide variety of scenes.

comment: 13 pages, 12 figures, v3: corrected typos in figures

♻ ☆ Beyond FVD: Enhanced Evaluation Metrics for Video Generation Quality

Ge Ya Luo, Gian Mario Favero, Zhi Hao Luo, Alexia Jolicoeur-Martineau, Christopher Pal

The Fr\'echet Video Distance (FVD) is a widely adopted metric for evaluating video generation distribution quality. However, its effectiveness relies on critical assumptions. Our analysis reveals three significant limitations: (1) the non-Gaussianity of the Inflated 3D Convnet (I3D) feature space; (2) the insensitivity of I3D features to temporal distortions; (3) the impractical sample sizes required for reliable estimation. These findings undermine FVD's reliability and show that FVD falls short as a standalone metric for video generation evaluation. After extensive analysis of a wide range of metrics and backbone architectures, we propose JEDi, the JEPA Embedding Distance, based on features derived from a Joint Embedding Predictive Architecture, measured using Maximum Mean Discrepancy with polynomial kernel. Our experiments on multiple open-source datasets show clear evidence that it is a superior alternative to the widely used FVD metric, requiring only 16% of the samples to reach its steady value, while increasing alignment with human evaluation by 34%, on average.

♻ ☆ MIBench: A Comprehensive Benchmark for Model Inversion Attack and Defense

Yixiang Qiu, Hongyao Yu, Hao Fang, Wenbo Yu, Bin Chen, Xuan Wang, Shu-Tao Xia, Ke Xu

Model Inversion (MI) attacks aim at leveraging the output information of target models to reconstruct privacy-sensitive training data, raising widespread concerns on privacy threats of Deep Neural Networks (DNNs). Unfortunately, in tandem with the rapid evolution of MI attacks, the lack of a comprehensive, aligned, and reliable benchmark has emerged as a formidable challenge. This deficiency leads to inadequate comparisons between different attack methods and inconsistent experimental setups. In this paper, we introduce the first practical benchmark for model inversion attacks and defenses to address this critical gap, which is named \textit{MIBench}. This benchmark serves as an extensible and reproducible modular-based toolbox and currently integrates a total of 16 state-of-the-art attack and defense methods. Moreover, we furnish a suite of assessment tools encompassing 9 commonly used evaluation protocols to facilitate standardized and fair evaluation and analysis. Capitalizing on this foundation, we conduct extensive experiments from multiple perspectives to holistically compare and analyze the performance of various methods across different scenarios, which overcomes the misalignment issues and discrepancy prevalent in previous works. Based on the collected attack methods and defense strategies, we analyze the impact of target resolution, defense robustness, model predictive power, model architectures, transferability and loss function. Our hope is that this \textit{MIBench} could provide a unified, practical and extensible toolbox and is widely utilized by researchers in the field to rigorously test and compare their novel methods, ensuring equitable evaluations and thereby propelling further advancements in the future development.

comment: 23 pages

♻ ☆ Control-oriented Clustering of Visual Latent Representation

Han Qi, Haocheng Yin, Heng Yang

We initiate a study of the geometry of the visual representation space -- the information channel from the vision encoder to the action decoder -- in an image-based control pipeline learned from behavior cloning. Inspired by the phenomenon of neural collapse (NC) in image classification, we investigate whether a similar law of clustering emerges in the visual representation space. Since image-based control is a regression task without explicitly defined classes, the central piece of the puzzle lies in determining according to what implicit classes the visual features cluster, if such a law exists. Focusing on image-based planar pushing, we posit the most important role of the visual representation in a control task is to convey a goal to the action decoder. We then classify training samples of expert demonstrations into eight "control-oriented" classes based on (a) the relative pose between the object and the target in the input or (b) the relative pose of the object induced by expert actions in the output, where one class corresponds to one relative pose orthant (REPO). Across four different instantiations of architecture, we report the prevalent emergence of control-oriented clustering in the visual representation space according to the eight REPOs. Beyond empirical observation, we show such a law of clustering can be leveraged as an algorithmic tool to improve test-time performance when training a policy with limited expert demonstrations. Particularly, we pretrain the vision encoder using NC as a regularization to encourage control-oriented clustering of the visual features. Surprisingly, such an NC-pretrained vision encoder, when finetuned end-to-end with the action decoder, boosts the test-time performance by 10% to 35% in the low-data regime. Real-world vision-based planar pushing experiments confirmed the surprising advantage of control-oriented visual representation pretraining.

♻ ☆ RoWeeder: Unsupervised Weed Mapping through Crop-Row Detection ECCV 2024

Pasquale De Marinis, Gennaro Vessio, Giovanna Castellano

Precision agriculture relies heavily on effective weed management to ensure robust crop yields. This study presents RoWeeder, an innovative framework for unsupervised weed mapping that combines crop-row detection with a noise-resilient deep learning model. By leveraging crop-row information to create a pseudo-ground truth, our method trains a lightweight deep learning model capable of distinguishing between crops and weeds, even in the presence of noisy data. Evaluated on the WeedMap dataset, RoWeeder achieves an F1 score of 75.3, outperforming several baselines. Comprehensive ablation studies further validated the model's performance. By integrating RoWeeder with drone technology, farmers can conduct real-time aerial surveys, enabling precise weed management across large fields. The code is available at: \url{https://github.com/pasqualedem/RoWeeder}.

comment: Computer Vision for Plant Phenotyping and Agriculture (CVPPA) workshop at ECCV 2024

♻ ☆ A Simple Image Segmentation Framework via In-Context Examples NeurIPS

Yang Liu, Chenchen Jing, Hengtao Li, Muzhi Zhu, Hao Chen, Xinlong Wang, Chunhua Shen

Recently, there have been explorations of generalist segmentation models that can effectively tackle a variety of image segmentation tasks within a unified in-context learning framework. However, these methods still struggle with task ambiguity in in-context segmentation, as not all in-context examples can accurately convey the task information. In order to address this issue, we present SINE, a simple image Segmentation framework utilizing in-context examples. Our approach leverages a Transformer encoder-decoder structure, where the encoder provides high-quality image representations, and the decoder is designed to yield multiple task-specific output masks to effectively eliminate task ambiguity. Specifically, we introduce an In-context Interaction module to complement in-context information and produce correlations between the target image and the in-context example and a Matching Transformer that uses fixed matching and a Hungarian algorithm to eliminate differences between different tasks. In addition, we have further perfected the current evaluation system for in-context image segmentation, aiming to facilitate a holistic appraisal of these models. Experiments on various segmentation tasks show the effectiveness of the proposed method.

comment: Accepted to Proc. Conference on Neural Information Processing Systems (NeurIPS) 2024. Webpage: https://github.com/aim-uofa/SINE

♻ ☆ iSeg: An Iterative Refinement-based Framework for Training-free Segmentation

Lin Sun, Jiale Cao, Jin Xie, Fahad Shahbaz Khan, Yanwei Pang

Stable diffusion has demonstrated strong image synthesis ability to given text descriptions, suggesting it to contain strong semantic clue for grouping objects. The researchers have explored employing stable diffusion for training-free segmentation. Most existing approaches refine cross-attention map by self-attention map once, demonstrating that self-attention map contains useful semantic information to improve segmentation. To fully utilize self-attention map, we present a deep experimental analysis on iteratively refining cross-attention map with self-attention map, and propose an effective iterative refinement framework for training-free segmentation, named iSeg. The proposed iSeg introduces an entropy-reduced self-attention module that utilizes a gradient descent scheme to reduce the entropy of self-attention map, thereby suppressing the weak responses corresponding to irrelevant global information. Leveraging the entropy-reduced self-attention module, our iSeg stably improves refined cross-attention map with iterative refinement. Further, we design a category-enhanced cross-attention module to generate accurate cross-attention map, providing a better initial input for iterative refinement. Extensive experiments across different datasets and diverse segmentation tasks reveal the merits of proposed contributions, leading to promising performance on diverse segmentation tasks. For unsupervised semantic segmentation on Cityscapes, our iSeg achieves an absolute gain of 3.8% in terms of mIoU compared to the best existing training-free approach in literature. Moreover, our proposed iSeg can support segmentation with different kinds of images and interactions. The project is available at https://linsun449.github.io/iSeg.

comment: Project Page: https://linsun449.github.io/iSeg/ Code: https://github.com/linsun449/iseg.code

♻ ☆ A Cat Is A Cat (Not A Dog!): Unraveling Information Mix-ups in Text-to-Image Encoders through Causal Analysis and Embedding Optimization NeurIPS 2024

Chieh-Yun Chen, Chiang Tseng, Li-Wu Tsao, Hong-Han Shuai

This paper analyzes the impact of causal manner in the text encoder of text-to-image (T2I) diffusion models, which can lead to information bias and loss. Previous works have focused on addressing the issues through the denoising process. However, there is no research discussing how text embedding contributes to T2I models, especially when generating more than one object. In this paper, we share a comprehensive analysis of text embedding: i) how text embedding contributes to the generated images and ii) why information gets lost and biases towards the first-mentioned object. Accordingly, we propose a simple but effective text embedding balance optimization method, which is training-free, with an improvement of 90.05% on information balance in stable diffusion. Furthermore, we propose a new automatic evaluation metric that quantifies information loss more accurately than existing methods, achieving 81% concordance with human assessments. This metric effectively measures the presence and accuracy of objects, addressing the limitations of current distribution scores like CLIP's text-image similarities.

comment: Accepted to NeurIPS 2024

Machine Learning 15

♻ ☆ Jogging the Memory of Unlearned LLMs Through Targeted Relearning Attacks

Shengyuan Hu, Yiwei Fu, Zhiwei Steven Wu, Virginia Smith

Machine unlearning is a promising approach to mitigate undesirable memorization of training data in LLMs. However, in this work we show that existing approaches for unlearning in LLMs are surprisingly susceptible to a simple set of targeted relearning attacks. With access to only a small and potentially loosely related set of data, we find that we can "jog" the memory of unlearned models to reverse the effects of unlearning. For example, we show that relearning on public medical articles can lead an unlearned LLM to output harmful knowledge about bioweapons, and relearning general wiki information about the book series Harry Potter can force the model to output verbatim memorized text. We formalize this unlearning-relearning pipeline, explore the attack across three popular unlearning benchmarks, and discuss future directions and guidelines that result from our study.

comment: 26 pages, 5 figures, 7 tables

♻ ☆ Beyond FVD: Enhanced Evaluation Metrics for Video Generation Quality

Ge Ya Luo, Gian Mario Favero, Zhi Hao Luo, Alexia Jolicoeur-Martineau, Christopher Pal

♻ ☆ A Simulation-Free Deep Learning Approach to Stochastic Optimal Control

Mengjian Hua, Matthieu Laurière, Eric Vanden-Eijnden

We propose a simulation-free algorithm for the solution of generic problems in stochastic optimal control (SOC). Unlike existing methods, our approach does not require the solution of an adjoint problem, but rather leverages Girsanov theorem to directly calculate the gradient of the SOC objective on-policy. This allows us to speed up the optimization of control policies parameterized by neural networks since it completely avoids the expensive back-propagation step through stochastic differential equations (SDEs) used in the Neural SDE framework. In particular, it enables us to solve SOC problems in high dimension and on long time horizons. We demonstrate the efficiency of our approach in various domains of applications, including standard stochastic optimal control problems, sampling from unnormalized distributions via construction of a Schr\"odinger-F\"ollmer process, and fine-tuning of pre-trained diffusion models. In all cases our method is shown to outperform the existing methods in both the computing time and memory efficiency.

♻ ☆ PAMLR: A Passive-Active Multi-Armed Bandit-Based Solution for LoRa Channel Allocation

Jihoon Yun, Chengzhang Li, Anish Arora

Achieving low duty cycle operation in low-power wireless networks in urban environments is complicated by the complex and variable dynamics of external interference and fading. We explore the use of reinforcement learning for achieving low power consumption for the task of optimal selection of channels. The learning relies on a hybrid of passive channel sampling for dealing with external interference and active channel sampling for dealing with fading. Our solution, Passive-Active Multi-armed bandit for LoRa (PAMLR, pronounced "Pamela"), balances the two types of samples to achieve energy-efficient channel selection: active channel measurements are tuned to an appropriately low level to update noise thresholds, and to compensate passive channel measurements are tuned to an appropriately high level for selecting the top-most channels from channel exploration using the noise thresholds. The rates of both types of samples are adapted in response to channel dynamics. Based on extensive testing in multiple environments in different cities, we validate that PAMLR can maintain excellent communication quality, as demonstrated by a low SNR regret compared to the optimal channel allocation policy, while substantially minimizing the energy cost associated with channel measurements.

comment: BuildSys 23

♻ ☆ SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Masks

Fenia Christopoulou, Ronald Cardenas, Gerasimos Lampouras, Haitham Bou-Ammar, Jun Wang

comment: 20 pages, 9 figures, 5 tables. Under Review

♻ ☆ Control-oriented Clustering of Visual Latent Representation

Han Qi, Haocheng Yin, Heng Yang

♻ ☆ FreSh: Frequency Shifting for Accelerated Neural Representation Learning

Adam Kania, Marko Mihajlovic, Sergey Prokudin, Jacek Tabor, Przemysław Spurek

Implicit Neural Representations (INRs) have recently gained attention as a powerful approach for continuously representing signals such as images, videos, and 3D shapes using multilayer perceptrons (MLPs). However, MLPs are known to exhibit a low-frequency bias, limiting their ability to capture high-frequency details accurately. This limitation is typically addressed by incorporating high-frequency input embeddings or specialized activation layers. In this work, we demonstrate that these embeddings and activations are often configured with hyperparameters that perform well on average but are suboptimal for specific input signals under consideration, necessitating a costly grid search to identify optimal settings. Our key observation is that the initial frequency spectrum of an untrained model's output correlates strongly with the model's eventual performance on a given target signal. Leveraging this insight, we propose frequency shifting (or FreSh), a method that selects embedding hyperparameters to align the frequency spectrum of the model's initial output with that of the target signal. We show that this simple initialization technique improves performance across various neural representation methods and tasks, achieving results comparable to extensive hyperparameter sweeps but with only marginal computational overhead compared to training a single model with default hyperparameters.

comment: Code at https://github.com/gmum/FreSh/

♻ ☆ DEPT: Decoupled Embeddings for Pre-training Language Models

Alex Iacob, Lorenzo Sani, Meghdad Kurmanji, William F. Shen, Xinchi Qiu, Dongqi Cai, Yan Gao, Nicholas D. Lane

♻ ☆ Neural Collaborative Filtering to Detect Anomalies in Human Semantic Trajectories SP

Yueyang Liu, Lance Kennedy, Hossein Amiri, Andreas Züfle

Human trajectory anomaly detection has become increasingly important across a wide range of applications, including security surveillance and public health. However, existing trajectory anomaly detection methods are primarily focused on vehicle-level traffic, while human-level trajectory anomaly detection remains under-explored. Since human trajectory data is often very sparse, machine learning methods have become the preferred approach for identifying complex patterns. However, concerns regarding potential biases and the robustness of these models have intensified the demand for more transparent and explainable alternatives. In response to these challenges, our research focuses on developing a lightweight anomaly detection model specifically designed to detect anomalies in human trajectories. We propose a Neural Collaborative Filtering approach to model and predict normal mobility. Our method is designed to model users' daily patterns of life without requiring prior knowledge, thereby enhancing performance in scenarios where data is sparse or incomplete, such as in cold start situations. Our algorithm consists of two main modules. The first is the collaborative filtering module, which applies collaborative filtering to model normal mobility of individual humans to places of interest. The second is the neural module, responsible for interpreting the complex spatio-temporal relationships inherent in human trajectory data. To validate our approach, we conducted extensive experiments using simulated and real-world datasets comparing to numerous state-of-the-art trajectory anomaly detection approaches.

comment: Accepted for publication in the 1st ACM SIGSPATIAL International Workshop on Geospatial Anomaly Detection (GeoAnomalies'24)

♻ ☆ MC-QDSNN: Quantized Deep evolutionary SNN with Multi-Dendritic Compartment Neurons for Stress Detection using Physiological Signals

Ajay B S, Phani Pavan K, Madhav Rao

Long short-term memory (LSTM) has emerged as a definitive network for analyzing and inferring time series data. LSTM has the capability to extract spectral features and a mixture of temporal features. Due to this benefit, a similar feature extraction method is explored for the spiking counterparts targeting time-series data. Though LSTMs perform well in their spiking form, they tend to be compute and power intensive. Addressing this issue, this work proposes Multi-Compartment Leaky (MCLeaky) neuron as a viable alternative for efficient processing of time series data. The MCLeaky neuron, derived from the Leaky Integrate and Fire (LIF) neuron model, contains multiple memristive synapses interlinked to form a memory component, which emulates the human brain's Hippocampus region. The proposed MCLeaky neuron based Spiking Neural Network model and its quantized variant were benchmarked against state-of-the-art (SOTA) Spiking LSTMs to perform human stress detection, by comparing compute requirements, latency and real-world performances on unseen data with models derived through Neural Architecture Search (NAS). Results show that networks with MCLeaky activation neuron managed a superior accuracy of 98.8% to detect stress based on Electrodermal Activity (EDA) signals, better than any other investigated models, while using 20% less parameters on average. MCLeaky neuron was also tested for various signals including EDA Wrist and Chest, Temperature, ECG, and combinations of them. Quantized MCLeaky model was also derived and validated to forecast their performance on hardware architectures, which resulted in 91.84% accuracy. The neurons were evaluated for multiple modalities of data towards stress detection, which resulted in energy savings of 25.12x to 39.20x and EDP gains of 52.37x to 81.9x over ANNs, while offering a best accuracy of 98.8% when compared with the rest of the SOTA implementations.

comment: 13 pages, 15 figures. Applied to IEEE Transactions on Computer Aided Design Journal. Awaiting a verdict

♻ ☆ Failure-Proof Non-Contrastive Self-Supervised Learning

Emanuele Sansone, Tim Lebailly, Tinne Tuytelaars

We identify sufficient conditions to avoid known failure modes, including representation, dimensional, cluster and intracluster collapses, occurring in non-contrastive self-supervised learning. Based on these findings, we propose a principled design for the projector and loss function. We theoretically demonstrate that this design introduces an inductive bias that promotes learning representations that are both decorrelated and clustered without explicit enforcing these properties and leading to improved generalization. To the best of our knowledge, this is the first solution that achieves robust training with respect to these failure modes while guaranteeing enhanced generalization performance in downstream tasks. We validate our theoretical findings on image datasets including SVHN, CIFAR10, CIFAR100 and ImageNet-100, and show that our solution, dubbed FALCON, outperforms existing feature decorrelation and cluster-based self-supervised learning methods in terms of generalization to clustering and linear classification tasks.

♻ ☆ Detecting and Approximating Redundant Computational Blocks in Neural Networks

Irene Cannistraci, Emanuele Rodolà, Bastian Rieck

Deep neural networks often learn similar internal representations, both across different models and within their own layers. While inter-network similarities have enabled techniques such as model stitching and merging, intra-network similarities present new opportunities for designing more efficient architectures. In this paper, we investigate the emergence of these internal similarities across different layers in diverse neural architectures, showing that similarity patterns emerge independently of the datataset used. We introduce a simple metric, Block Redundancy, to detect redundant blocks, providing a foundation for future architectural optimization methods. Building on this, we propose Redundant Blocks Approximation (RBA), a general framework that identifies and approximates one or more redundant computational blocks using simpler transformations. We show that the transformation $\mathcal{T}$ between two representations can be efficiently computed in closed-form, and it is enough to replace the redundant blocks from the network. RBA reduces model parameters and time complexity while maintaining good performance. We validate our method on classification tasks in the vision domain using a variety of pretrained foundational models and datasets.

comment: 9 pages, 10 figures, 7 tables

♻ ☆ Goal-Conditioned Terminal Value Estimation for Real-time and Multi-task Model Predictive Control

Mitsuki Morita, Satoshi Yamamori, Satoshi Yagi, Norikazu Sugimoto, Jun Morimoto

While MPC enables nonlinear feedback control by solving an optimal control problem at each timestep, the computational burden tends to be significantly large, making it difficult to optimize a policy within the control period. To address this issue, one possible approach is to utilize terminal value learning to reduce computational costs. However, the learned value cannot be used for other tasks in situations where the task dynamically changes in the original MPC setup. In this study, we develop an MPC framework with goal-conditioned terminal value learning to achieve multitask policy optimization while reducing computational time. Furthermore, by using a hierarchical control structure that allows the upper-level trajectory planner to output appropriate goal-conditioned trajectories, we demonstrate that a robot model is able to generate diverse motions. We evaluate the proposed method on a bipedal inverted pendulum robot model and confirm that combining goal-conditioned terminal value learning with an upper-level trajectory planner enables real-time control; thus, the robot successfully tracks a target trajectory on sloped terrain.

comment: 16 pages, 9 figures

♻ ☆ OptiBench Meets ReSocratic: Measure and Improve LLMs for Optimization Modeling

Zhicheng Yang, Yiwei Wang, Yinya Huang, Zhijiang Guo, Wei Shi, Xiongwei Han, Liang Feng, Linqi Song, Xiaodan Liang, Jing Tang

Large language models (LLMs) have exhibited their problem-solving abilities in mathematical reasoning. Solving realistic optimization (OPT) problems in application scenarios requires advanced and applied mathematics ability. However, current OPT benchmarks that merely solve linear programming are far from complex realistic situations. In this work, we propose OptiBench, a benchmark for End-to-end optimization problem-solving with human-readable inputs and outputs. OptiBench contains rich optimization problems, including linear and nonlinear programming with or without tabular data, which can comprehensively evaluate LLMs' solving ability. In our benchmark, LLMs are required to call a code solver to provide precise numerical answers. Furthermore, to alleviate the data scarcity for optimization problems, and to bridge the gap between open-source LLMs on a small scale (e.g., Llama-3-8b) and closed-source LLMs (e.g., GPT-4), we further propose a data synthesis method namely ReSocratic. Unlike general data synthesis methods that proceed from questions to answers, \ReSocratic first incrementally synthesizes formatted optimization demonstration with mathematical formulations step by step and then back-translates the generated demonstrations into questions. Based on this, we synthesize the ReSocratic-29k dataset. We further conduct supervised fine-tuning with ReSocratic-29k on multiple open-source models. Experimental results show that ReSocratic-29k significantly improves the performance of open-source models.

♻ ☆ Strong Model Collapse

Elvis Dohmatob, Yunzhen Feng, Arjun Subramonian, Julia Kempe

Within the scaling laws paradigm, which underpins the training of large neural networks like ChatGPT and Llama, we consider a supervised regression setting and establish the existance of a strong form of the model collapse phenomenon, a critical performance degradation due to synthetic data in the training corpus. Our results show that even the smallest fraction of synthetic data (e.g., as little as 1\% of the total training dataset) can still lead to model collapse: larger and larger training sets do not enhance performance. We further investigate whether increasing model size, an approach aligned with current trends in training large language models, exacerbates or mitigates model collapse. In a simplified regime where neural networks are approximated via random projections of tunable size, we both theoretically and empirically show that larger models can amplify model collapse. Interestingly, our theory also indicates that, beyond the interpolation threshold (which can be extremely high for very large datasets), larger models may mitigate the collapse, although they do not entirely prevent it. Our theoretical findings are empirically verified through experiments on language models and feed-forward neural networks for images.

2024-10-07

Programming Languages 3

♻ ☆ Solving the Phase Ordering Problem $\ne$ Generating the Globally Optimal Code

Yu Wang, Hongyu Chen, Ke Wang

Phase ordering problem has been a long-standing challenge in compiler optimizations. Over the past four decades, a significant amount of effort has been devoted, and indeed, substantial progress has been made. However, in this paper, we raise questions about the overall significance of solving the phase ordering problem in the first place, as pursuing a solution to this problem may not align with the fundamental goal of compiler optimizations, i.e., generating the globally optimal code among all programs that compilers deem semantically equivalent to an input program. Our findings, supported by both theoretical and empirical evidence, show that solving the phase ordering problem is not equivalent to generating such globally optimal code. The fundamental reason that applying the optimal phase ordering may still result in suboptimal code is the exclusion of programs of less efficiency during the optimization process. Motivated by this insight, we propose a theoretical approach, called \textit{infinitive iterative bi-directional optimizations} (\textit{IIBO}), which is guaranteed to converge to the globally optimal code for any input program. We realize IIBO into a practical algorithm and apply it to optimize real-world programs. Results show that IIBO frequently generates more efficient code than GCC/LLVM, two state-of-the-art industry compilers, as well as exhaustive search, which can be deemed the solution to the phasing ordering problem.% input programs. Given the significance and impact of our results, we are currently in active discussions with LLVM engineers on the possible incorporation of our findings into their next release. In general, we expect our work to inspire new design principles for compiler development in the pursuit of generating the globally optimal code.

♻ ☆ A Calculus for Scoped Effects & Handlers

Roger Bosman, Birthe van den Berg, Wenhao Tang, Tom Schrijvers

Algebraic effects & handlers have become a standard approach for side-effects in functional programming. Their modular composition with other effects and clean separation of syntax and semantics make them attractive to a wide audience. However, not all effects can be classified as algebraic; some need a more sophisticated handling. In particular, effects that have or create a delimited scope need special care, as their continuation consists of two parts-in and out of the scope-and their modular composition introduces additional complexity. These effects are called scoped and have gained attention by their growing applicability and adoption in popular libraries. While calculi have been designed with algebraic effects & handlers built in to facilitate their use, a calculus that supports scoped effects & handlers in a similar manner does not yet exist. This work fills this gap: we present $\lambda_{\mathit{sc}}$, a calculus with native support for both algebraic and scoped effects & handlers. It addresses the need for polymorphic handlers and explicit clauses for forwarding unknown scoped operations to other handlers. Our calculus is based on Eff, an existing calculus for algebraic effects, extended with Koka-style row polymorphism, and consists of a formal grammar, operational semantics, a (type-safe) type-and-effect system and type inference. We demonstrate $\lambda_{\mathit{sc}}$ on a range of examples.

♻ ☆ RTLCoder: Fully Open-Source and Efficient LLM-Assisted RTL Code Generation Technique

Shang Liu, Wenji Fang, Yao Lu, Jing Wang, Qijun Zhang, Hongce Zhang, Zhiyao Xie

The automatic generation of RTL code (e.g., Verilog) using natural language instructions and large language models (LLMs) has attracted significant research interest recently. However, most existing approaches heavily rely on commercial LLMs such as ChatGPT, while open-source LLMs tailored for this specific design generation task exhibit notably inferior performance. The absence of high-quality open-source solutions restricts the flexibility and data privacy of this emerging technique. In this study, we present a new customized LLM solution with a modest parameter count of only 7B, achieving better performance than GPT-3.5 on all representative benchmarks for RTL code generation. Especially, it outperforms GPT-4 in VerilogEval Machine benchmark. This remarkable balance between accuracy and efficiency is made possible by leveraging our new RTL code dataset and a customized LLM algorithm, both of which have been made fully open-source. Furthermore, we have successfully quantized our LLM to 4-bit with a total size of 4GB, enabling it to function on a single laptop with only slight performance degradation. This efficiency allows the RTL generator to serve as a local assistant for engineers, ensuring all design privacy concerns are addressed.

comment: Accepted by IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Computation and Language 142

☆ Data Advisor: Dynamic Data Curation for Safety Alignment of Large Language Models EMNLP 2024

Fei Wang, Ninareh Mehrabi, Palash Goyal, Rahul Gupta, Kai-Wei Chang, Aram Galstyan

Data is a crucial element in large language model (LLM) alignment. Recent studies have explored using LLMs for efficient data collection. However, LLM-generated data often suffers from quality issues, with underrepresented or absent aspects and low-quality datapoints. To address these problems, we propose Data Advisor, an enhanced LLM-based method for generating data that takes into account the characteristics of the desired dataset. Starting from a set of pre-defined principles in hand, Data Advisor monitors the status of the generated data, identifies weaknesses in the current dataset, and advises the next iteration of data generation accordingly. Data Advisor can be easily integrated into existing data generation methods to enhance data quality and coverage. Experiments on safety alignment of three representative LLMs (i.e., Mistral, Llama2, and Falcon) demonstrate the effectiveness of Data Advisor in enhancing model safety against various fine-grained safety issues without sacrificing model utility.

comment: Accepted to EMNLP 2024 Main Conference. Project website: https://feiwang96.github.io/DataAdvisor/

☆ Grounding Partially-Defined Events in Multimodal Data EMNLP

Kate Sanders, Reno Kriz, David Etter, Hannah Recknor, Alexander Martin, Cameron Carpenter, Jingyang Lin, Benjamin Van Durme

How are we able to learn about complex current events just from short snippets of video? While natural language enables straightforward ways to represent under-specified, partially observable events, visual data does not facilitate analogous methods and, consequently, introduces unique challenges in event understanding. With the growing prevalence of vision-capable AI agents, these systems must be able to model events from collections of unstructured video data. To tackle robust event modeling in multimodal settings, we introduce a multimodal formulation for partially-defined events and cast the extraction of these events as a three-stage span retrieval task. We propose a corresponding benchmark for this task, MultiVENT-G, that consists of 14.5 hours of densely annotated current event videos and 1,168 text documents, containing 22.8K labeled event-centric entities. We propose a collection of LLM-driven approaches to the task of multimodal event analysis, and evaluate them on MultiVENT-G. Results illustrate the challenges that abstract event understanding poses and demonstrates promise in event-centric video-language systems.

comment: Preprint; 9 pages; 2024 EMNLP Findings

☆ PrefixQuant: Static Quantization Beats Dynamic through Prefixed Outliers in LLMs

Mengzhao Chen, Yi Liu, Jiahao Wang, Yi Bin, Wenqi Shao, Ping Luo

Quantization is essential for deploying Large Language Models (LLMs) by enhancing memory efficiency and inference speed. Existing methods for activation quantization mainly address channel-wise outliers, often neglecting token-wise outliers, leading to reliance on costly per-token dynamic quantization. To address this, we introduce PrefixQuant, a novel technique that isolates outlier tokens offline without re-training. Specifically, PrefixQuant identifies high-frequency outlier tokens and prefixes them in the KV cache, preventing the generation of outlier tokens during inference and simplifying quantization. To our knowledge, PrefixQuant is the first to enable efficient per-tensor static quantization to outperform expensive per-token dynamic quantization. For instance, in W4A4KV4 (4- bit weight, 4-bit activation, and 4-bit KV cache) Llama-3-8B, PrefixQuant with per-tensor static quantization achieves a 7.43 WikiText2 perplexity and 71.08% average accuracy on 5 common-sense reasoning tasks, outperforming previous per-token dynamic quantization methods like QuaRot with 0.98 perplexity improvement and +5.98 points accuracy. Additionally, the inference speed of W4A4 quantized models using PrefixQuant is 1.60x to 2.81x faster than FP16 models and exceeds QuaRot models by 1.2x to 1.3x. Our code is available at \url{https://github.com/ChenMnZ/PrefixQuant}.

comment: A PTQ method to significantly boost the performance of static activation quantization

☆ TurtleBench: Evaluating Top Language Models via Real-World Yes/No Puzzles

Qingchen Yu, Shichao Song, Ke Fang, Yunfeng Shi, Zifan Zheng, Hanyu Wang, Simin Niu, Zhiyu Li

As the application of Large Language Models (LLMs) expands, the demand for reliable evaluations increases. Existing LLM evaluation benchmarks primarily rely on static datasets, making it challenging to assess model performance in dynamic interactions with users. Moreover, these benchmarks often depend on specific background knowledge, complicating the measurement of a model's logical reasoning capabilities. Other dynamic evaluation methods based on strong models or manual efforts may introduce biases and incur high costs and time demands, hindering large-scale application. To address these issues, we propose TurtleBench. TurtleBench collects real user guesses from our online Turtle Soup Puzzle platform that we developed. This approach allows for the relatively dynamic generation of evaluation datasets, mitigating the risk of model cheating while aligning assessments more closely with genuine user needs for reasoning capabilities, thus enhancing the reliability of evaluations. TurtleBench includes 1,532 user guesses along with the correctness of guesses after annotation. Using this dataset, we thoroughly evaluated nine of the most advanced LLMs available today. Notably, the OpenAI o1 series models did not achieve leading results in these evaluations. We propose several hypotheses for further research, such as "the latent reasoning of o1 utilizes trivial Chain-of-Thought (CoT) techniques" and "increasing CoT length not only provides reasoning benefits but also incurs noise costs."

comment: 22 pages

☆ Differential Transformer

Tianzhu Ye, Li Dong, Yuqing Xia, Yutao Sun, Yi Zhu, Gao Huang, Furu Wei

Transformer tends to overallocate attention to irrelevant context. In this work, we introduce Diff Transformer, which amplifies attention to the relevant context while canceling noise. Specifically, the differential attention mechanism calculates attention scores as the difference between two separate softmax attention maps. The subtraction cancels noise, promoting the emergence of sparse attention patterns. Experimental results on language modeling show that Diff Transformer outperforms Transformer in various settings of scaling up model size and training tokens. More intriguingly, it offers notable advantages in practical applications, such as long-context modeling, key information retrieval, hallucination mitigation, in-context learning, and reduction of activation outliers. By being less distracted by irrelevant context, Diff Transformer can mitigate hallucination in question answering and text summarization. For in-context learning, Diff Transformer not only enhances accuracy but is also more robust to order permutation, which was considered as a chronic robustness issue. The results position Diff Transformer as a highly effective and promising architecture to advance large language models.

☆ GLEE: A Unified Framework and Benchmark for Language-based Economic Environments

Eilam Shapira, Omer Madmon, Itamar Reinman, Samuel Joseph Amouyal, Roi Reichart, Moshe Tennenholtz

Large Language Models (LLMs) show significant potential in economic and strategic interactions, where communication via natural language is often prevalent. This raises key questions: Do LLMs behave rationally? Can they mimic human behavior? Do they tend to reach an efficient and fair outcome? What is the role of natural language in the strategic interaction? How do characteristics of the economic environment influence these dynamics? These questions become crucial concerning the economic and societal implications of integrating LLM-based agents into real-world data-driven systems, such as online retail platforms and recommender systems. While the ML community has been exploring the potential of LLMs in such multi-agent setups, varying assumptions, design choices and evaluation criteria across studies make it difficult to draw robust and meaningful conclusions. To address this, we introduce a benchmark for standardizing research on two-player, sequential, language-based games. Inspired by the economic literature, we define three base families of games with consistent parameterization, degrees of freedom and economic measures to evaluate agents' performance (self-gain), as well as the game outcome (efficiency and fairness). We develop an open-source framework for interaction simulation and analysis, and utilize it to collect a dataset of LLM vs. LLM interactions across numerous game configurations and an additional dataset of human vs. LLM interactions. Through extensive experimentation, we demonstrate how our framework and dataset can be used to: (i) compare the behavior of LLM-based agents to human players in various economic contexts; (ii) evaluate agents in both individual and collective performance measures; and (iii) quantify the effect of the economic characteristics of the environments on the behavior of agents.

☆ Causal Micro-Narratives EMNLP 2024

Mourad Heddaya, Qingcheng Zeng, Chenhao Tan, Rob Voigt, Alexander Zentefis

We present a novel approach to classify causal micro-narratives from text. These narratives are sentence-level explanations of the cause(s) and/or effect(s) of a target subject. The approach requires only a subject-specific ontology of causes and effects, and we demonstrate it with an application to inflation narratives. Using a human-annotated dataset spanning historical and contemporary US news articles for training, we evaluate several large language models (LLMs) on this multi-label classification task. The best-performing model--a fine-tuned Llama 3.1 8B--achieves F1 scores of 0.87 on narrative detection and 0.71 on narrative classification. Comprehensive error analysis reveals challenges arising from linguistic ambiguity and highlights how model errors often mirror human annotator disagreements. This research establishes a framework for extracting causal micro-narratives from real-world data, with wide-ranging applications to social science research.

comment: Accepted to EMNLP 2024 Workshop on Narrative Understanding

☆ SFTMix: Elevating Language Model Instruction Tuning with Mixup Recipe

Yuxin Xiao, Shujian Zhang, Wenxuan Zhou, Marzyeh Ghassemi, Sanqiang Zhao

To induce desired behaviors in large language models (LLMs) for interaction-driven tasks, the instruction-tuning stage typically trains LLMs on instruction-response pairs using the next-token prediction (NTP) loss. Previous work aiming to improve instruction-tuning performance often emphasizes the need for higher-quality supervised fine-tuning (SFT) datasets, which typically involves expensive data filtering with proprietary LLMs or labor-intensive data generation by human annotators. However, these approaches do not fully leverage the datasets' intrinsic properties, resulting in high computational and labor costs, thereby limiting scalability and performance gains. In this paper, we propose SFTMix, a novel recipe that elevates instruction-tuning performance beyond the conventional NTP paradigm, without the need for well-curated datasets. Observing that LLMs exhibit uneven confidence across the semantic representation space, we argue that examples with different confidence levels should play distinct roles during the instruction-tuning process. Based on this insight, SFTMix leverages training dynamics to identify examples with varying confidence levels, then applies a Mixup-based regularization to mitigate overfitting on confident examples while propagating supervision signals to improve learning on relatively unconfident ones. This approach enables SFTMix to significantly outperform NTP across a wide range of instruction-following and healthcare domain-specific SFT tasks, demonstrating its adaptability to diverse LLM families and scalability to datasets of any size. Comprehensive ablation studies further verify the robustness of SFTMix's design choices, underscoring its versatility in consistently enhancing performance across different LLMs and datasets in broader natural language processing applications.

☆ Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents

Boyu Gou, Ruohan Wang, Boyuan Zheng, Yanan Xie, Cheng Chang, Yiheng Shu, Huan Sun, Yu Su

Multimodal large language models (MLLMs) are transforming the capabilities of graphical user interface (GUI) agents, facilitating their transition from controlled simulations to complex, real-world applications across various platforms. However, the effectiveness of these agents hinges on the robustness of their grounding capability. Current GUI agents predominantly utilize text-based representations such as HTML or accessibility trees, which, despite their utility, often introduce noise, incompleteness, and increased computational overhead. In this paper, we advocate a human-like embodiment for GUI agents that perceive the environment entirely visually and directly take pixel-level operations on the GUI. The key is visual grounding models that can accurately map diverse referring expressions of GUI elements to their coordinates on the GUI across different platforms. We show that a simple recipe, which includes web-based synthetic data and slight adaptation of the LLaVA architecture, is surprisingly effective for training such visual grounding models. We collect the largest dataset for GUI visual grounding so far, containing 10M GUI elements and their referring expressions over 1.3M screenshots, and use it to train UGround, a strong universal visual grounding model for GUI agents. Empirical results on six benchmarks spanning three categories (grounding, offline agent, and online agent) show that 1) UGround substantially outperforms existing visual grounding models for GUI agents, by up to 20% absolute, and 2) agents with UGround outperform state-of-the-art agents, despite the fact that existing agents use additional text-based input while ours only uses visual perception. These results provide strong support for the feasibility and promises of GUI agents that navigate the digital world as humans do.

☆ TuneVLSeg: Prompt Tuning Benchmark for Vision-Language Segmentation Models ACCV 2024

Rabin Adhikari, Safal Thapaliya, Manish Dhakal, Bishesh Khanal

comment: Accepted at ACCV 2024 (oral presentation)

☆ CasiMedicos-Arg: A Medical Question Answering Dataset Annotated with Explanatory Argumentative Structures

katerina Sviridova, Anar Yeginbergen, Ainara Estarrona, Elena Cabrio, Serena Villata, Rodrigo Agerri

comment: 9 pages

☆ Cookbook: A framework for improving LLM generative abilities via programmatic data generating templates

Avanika Narayan, Mayee F. Chen, Kush Bhatia, Christopher Ré

Fine-tuning large language models (LLMs) on instruction datasets is a common way to improve their generative capabilities. However, instruction datasets can be expensive and time-consuming to manually curate, and while LLM-generated data is less labor-intensive, it may violate user privacy agreements or terms of service of LLM providers. Therefore, we seek a way of constructing instruction datasets with samples that are not generated by humans or LLMs but still improve LLM generative capabilities. In this work, we introduce Cookbook, a framework that programmatically generates training data consisting of simple patterns over random tokens, resulting in a scalable, cost-effective approach that avoids legal and privacy issues. First, Cookbook uses a template -- a data generating Python function -- to produce training data that encourages the model to learn an explicit pattern-based rule that corresponds to a desired task. We find that fine-tuning on Cookbook-generated data is able to improve performance on its corresponding task by up to 52.7 accuracy points. Second, since instruction datasets improve performance on multiple downstream tasks simultaneously, Cookbook algorithmically learns how to mix data from various templates to optimize performance on multiple tasks. On the standard multi-task GPT4ALL evaluation suite, Mistral-7B fine-tuned using a Cookbook-generated dataset attains the best accuracy on average compared to other 7B parameter instruction-tuned models and is the best performing model on 3 out of 8 tasks. Finally, we analyze when and why Cookbook improves performance and present a metric that allows us to verify that the improvement is largely explained by the model's generations adhering better to template rules.

comment: COLM 2024

☆ Precise Model Benchmarking with Only a Few Observations EMNLP 2024

Riccardo Fogliato, Pratik Patil, Nil-Jana Akpinar, Mathew Monfort

How can we precisely estimate a large language model's (LLM) accuracy on questions belonging to a specific topic within a larger question-answering dataset? The standard direct estimator, which averages the model's accuracy on the questions in each subgroup, may exhibit high variance for subgroups (topics) with small sample sizes. Synthetic regression modeling, which leverages the model's accuracy on questions about other topics, may yield biased estimates that are too unreliable for large subgroups. We prescribe a simple yet effective solution: an empirical Bayes (EB) estimator that balances direct and regression estimates for each subgroup separately, improving the precision of subgroup-level estimates of model performance. Our experiments on multiple datasets show that this approach consistently provides more precise estimates of the LLM performance compared to the direct and regression approaches, achieving substantial reductions in the mean squared error. Confidence intervals for EB estimates also have near-nominal coverage and are narrower compared to those for the direct estimator. Additional experiments on tabular and vision data validate the benefits of this EB approach.

comment: To appear at EMNLP 2024

☆ Density estimation with LLMs: a geometric investigation of in-context learning trajectories ICLR 2025

Toni J. B. Liu, Nicolas Boullé, Raphaël Sarfati, Christopher J. Earls

Large language models (LLMs) demonstrate remarkable emergent abilities to perform in-context learning across various tasks, including time series forecasting. This work investigates LLMs' ability to estimate probability density functions (PDFs) from data observed in-context; such density estimation (DE) is a fundamental task underlying many probabilistic modeling problems. We leverage the Intensive Principal Component Analysis (InPCA) to visualize and analyze the in-context learning dynamics of LLaMA-2 models. Our main finding is that these LLMs all follow similar learning trajectories in a low-dimensional InPCA space, which are distinct from those of traditional density estimation methods like histograms and Gaussian kernel density estimation (KDE). We interpret the LLaMA in-context DE process as a KDE with an adaptive kernel width and shape. This custom kernel model captures a significant portion of LLaMA's behavior despite having only two parameters. We further speculate on why LLaMA's kernel width and shape differs from classical algorithms, providing insights into the mechanism of in-context probabilistic reasoning in LLMs.

comment: Under review as a conference paper at ICLR 2025

☆ Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality EMNLP 2024

Youngtaek Oh, Jae Won Cho, Dong-Jin Kim, In So Kweon, Junmo Kim

In this paper, we propose a new method to enhance compositional understanding in pre-trained vision and language models (VLMs) without sacrificing performance in zero-shot multi-modal tasks. Traditional fine-tuning approaches often improve compositional reasoning at the cost of degrading multi-modal capabilities, primarily due to the use of global hard negative (HN) loss, which contrasts global representations of images and texts. This global HN loss pushes HN texts that are highly similar to the original ones, damaging the model's multi-modal representations. To overcome this limitation, we propose Fine-grained Selective Calibrated CLIP (FSC-CLIP), which integrates local hard negative loss and selective calibrated regularization. These innovations provide fine-grained negative supervision while preserving the model's representational integrity. Our extensive evaluations across diverse benchmarks for both compositionality and multi-modal tasks show that FSC-CLIP not only achieves compositionality on par with state-of-the-art models but also retains strong multi-modal capabilities. Code is available at: https://github.com/ytaek-oh/fsc-clip.

comment: EMNLP 2024 (Long, Main). Project page: https://ytaek-oh.github.io/fsc-clip

☆ Studying and Mitigating Biases in Sign Language Understanding Models

Katherine Atwell, Danielle Bragg, Malihe Alikhani

Ensuring that the benefits of sign language technologies are distributed equitably among all community members is crucial. Thus, it is important to address potential biases and inequities that may arise from the design or use of these resources. Crowd-sourced sign language datasets, such as the ASL Citizen dataset, are great resources for improving accessibility and preserving linguistic diversity, but they must be used thoughtfully to avoid reinforcing existing biases. In this work, we utilize the rich information about participant demographics and lexical features present in the ASL Citizen dataset to study and document the biases that may result from models trained on crowd-sourced sign datasets. Further, we apply several bias mitigation techniques during model training, and find that these techniques reduce performance disparities without decreasing accuracy. With the publication of this work, we release the demographic information about the participants in the ASL Citizen dataset to encourage future bias mitigation work in this space.

☆ RevisEval: Improving LLM-as-a-Judge via Response-Adapted References

Qiyuan Zhang, Yufei Wang, Tiezheng YU, Yuxin Jiang, Chuhan Wu, Liangyou Li, Yasheng Wang, Xin Jiang, Lifeng Shang, Ruiming Tang, Fuyuan Lyu, Chen Ma

With significant efforts in recent studies, LLM-as-a-Judge has become a cost-effective alternative to human evaluation for assessing the text generation quality in a wide range of tasks. However, there still remains a reliability gap between LLM-as-a-Judge and human evaluation. One important reason is the lack of guided oracles in the evaluation process. Motivated by the role of reference pervasively used in classic text evaluation, we introduce RevisEval, a novel text generation evaluation paradigm via the response-adapted references. RevisEval is driven by the key observation that an ideal reference should maintain the necessary relevance to the response to be evaluated. Specifically, RevisEval leverages the text revision capabilities of large language models (LLMs) to adaptively revise the response, then treat the revised text as the reference (response-adapted reference) for the subsequent evaluation. Extensive experiments demonstrate that RevisEval outperforms traditional reference-free and reference-based evaluation paradigms that use LLM-as-a-Judge across NLG tasks and open-ended instruction-following tasks. More importantly, our response-adapted references can further boost the classical text metrics, e.g., BLEU and BERTScore, compared to traditional references and even rival the LLM-as-a-Judge. A detailed analysis is also conducted to confirm RevisEval's effectiveness in bias reduction, the impact of inference cost, and reference relevance.

☆ Understanding Warmup-Stable-Decay Learning Rates: A River Valley Loss Landscape Perspective

Kaiyue Wen, Zhiyuan Li, Jason Wang, David Hall, Percy Liang, Tengyu Ma

Training language models currently requires pre-determining a fixed compute budget because the typical cosine learning rate schedule depends on the total number of steps. In contrast, the Warmup-Stable-Decay (WSD) schedule uses a constant learning rate to produce a main branch of iterates that can in principle continue indefinitely without a pre-specified compute budget. Then, given any compute budget, one can branch out from the main branch at a proper at any time with a rapidly decaying learning rate to produce a strong model. Empirically, WSD generates a non-traditional loss curve: the loss remains elevated during the stable phase but sharply declines during the decay phase. Towards explaining this phenomenon, we conjecture that pretraining loss exhibits a river valley landscape, which resembles a deep valley with a river at its bottom. Under this assumption, we show that during the stable phase, the iterate undergoes large oscillations due to the high learning rate, yet it progresses swiftly along the river. During the decay phase, the rapidly dropping learning rate minimizes the iterate's oscillations, moving it closer to the river and revealing true optimization progress. Therefore, the sustained high learning rate phase and fast decaying phase are responsible for progress in the river and the mountain directions respectively, and are both critical. Our analysis predicts phenomenons consistent with empirical observations and shows that this landscape can emerge from pretraining on a simple bi-gram dataset. Inspired by the theory, we introduce WSD-S, a variant of WSD that reuses previous checkpoints' decay phases and keeps only one main branch, where we resume from a decayed checkpoint. WSD-S empirically outperforms WSD and Cyclic-Cosine in obtaining multiple language model checkpoints across various compute budgets in a single run for parameters scaling from 0.1B to 1.2B.

comment: 45 pages,13 figures

☆ Beyond Correlation: Interpretable Evaluation of Machine Translation Metrics EMNLP 2024

Stefano Perrella, Lorenzo Proietti, Pere-Lluís Huguet Cabot, Edoardo Barba, Roberto Navigli

Machine Translation (MT) evaluation metrics assess translation quality automatically. Recently, researchers have employed MT metrics for various new use cases, such as data filtering and translation re-ranking. However, most MT metrics return assessments as scalar scores that are difficult to interpret, posing a challenge to making informed design choices. Moreover, MT metrics' capabilities have historically been evaluated using correlation with human judgment, which, despite its efficacy, falls short of providing intuitive insights into metric performance, especially in terms of new metric use cases. To address these issues, we introduce an interpretable evaluation framework for MT metrics. Within this framework, we evaluate metrics in two scenarios that serve as proxies for the data filtering and translation re-ranking use cases. Furthermore, by measuring the performance of MT metrics using Precision, Recall, and F-score, we offer clearer insights into their capabilities than correlation with human judgments. Finally, we raise concerns regarding the reliability of manually curated data following the Direct Assessments+Scalar Quality Metrics (DA+SQM) guidelines, reporting a notably low agreement with Multidimensional Quality Metrics (MQM) annotations.

comment: Accepted at EMNLP 2024 Main Conference. 26 pages

☆ Enhancing Equity in Large Language Models for Medical Applications

Yuelyu Ji, Wenhe Ma, Sonish Sivarajkumar, Hang Zhang, Eugene Mathew Sadhu, Zhuochun Li, Xizhi Wu, Shyam Visweswaran, Yanshan Wang

Recent advancements have highlighted the potential of large language models (LLMs) in medical applications, notably in automating Clinical Trial Matching for translational research and providing medical question-answering for clinical decision support. However, our study reveals significant inequities in the use of LLMs, particularly for individuals from specific racial, gender, and underrepresented groups influenced by social determinants of health. These disparities could worsen existing health inequities if LLMs are broadly adopted in healthcare. To address this, we propose and evaluate a novel framework, EquityGuard, designed to detect and mitigate biases in LLM-based medical applications. EquityGuard incorporates a Bias Detection Mechanism capable of identifying and correcting unfair predictions, thus enhancing outcomes and promoting equity across diverse population groups.

☆ ReasoningRank: Teaching Student Models to Rank through Reasoning-Based Knowledge Distillation

Yuelyu Ji, Zhuochun Li, Rui Meng, Daqing He

Reranking documents based on their relevance to a given query is critical in information retrieval. Traditional reranking methods often focus on improving the initial rankings but lack transparency, failing to explain why one document is ranked higher. In this paper, we introduce ReasoningRank, a novel reranking approach that enhances clarity by generating two types of reasoning: explicit reasoning, which explains how a document addresses the query, and comparison reasoning, which justifies the relevance of one document over another. We leverage large language models (LLMs) as teacher models to generate these explanations and distill this knowledge into smaller, more resource-efficient student models. While the student models may not outperform LLMs in speed, they significantly reduce the computational burden by requiring fewer resources, making them more suitable for large-scale or resource-constrained settings. These student models are trained to both generate meaningful reasoning and rerank documents, achieving competitive performance across multiple datasets, including MSMARCO and BRIGHT. Experiments demonstrate that ReasoningRank improves reranking accuracy and provides valuable insights into the decision-making process, offering a structured and interpretable solution for reranking tasks.

☆ Efficient Inference for Large Language Model-based Generative Recommendation

Xinyu Lin, Chaoqun Yang, Wenjie Wang, Yongqi Li, Cunxiao Du, Fuli Feng, See-Kiong Ng, Tat-Seng Chua

☆ Deciphering the Interplay of Parametric and Non-parametric Memory in Retrieval-augmented Language Models EMNLP 2024

Mehrdad Farahani, Richard Johansson

Generative language models often struggle with specialized or less-discussed knowledge. A potential solution is found in Retrieval-Augmented Generation (RAG) models which act like retrieving information before generating responses. In this study, we explore how the \textsc{Atlas} approach, a RAG model, decides between what it already knows (parametric) and what it retrieves (non-parametric). We use causal mediation analysis and controlled experiments to examine how internal representations influence information processing. Our findings disentangle the effects of parametric knowledge and the retrieved context. They indicate that in cases where the model can choose between both types of information (parametric and non-parametric), it relies more on the context than the parametric knowledge. Furthermore, the analysis investigates the computations involved in \emph{how} the model uses the information from the context. We find that multiple mechanisms are active within the model and can be detected with mediation analysis: first, the decision of \emph{whether the context is relevant}, and second, how the encoder computes output representations to support copying when relevant.

comment: Accepted at EMNLP 2024

☆ VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks

Ziyan Jiang, Rui Meng, Xinyi Yang, Semih Yavuz, Yingbo Zhou, Wenhu Chen

Embedding models have been crucial in enabling various downstream tasks such as semantic similarity, information retrieval, and clustering. Recently, there has been a surge of interest in developing universal text embedding models that can generalize across tasks (e.g., MTEB). However, progress in learning universal multimodal embedding models has been relatively slow despite their importance. In this work, we aim to explore the potential for building universal embeddings capable of handling a wide range of downstream tasks. Our contributions are twofold: (1) MMEB (Massive Multimodal Embedding Benchmark), which covers 4 meta-tasks (i.e. classification, visual question answering, multimodal retrieval, and visual grounding) and 36 datasets, including 20 training and 16 evaluation datasets, and (2) VLM2Vec (Vision-Language Model -> Vector), a contrastive training framework that converts any state-of-the-art vision-language model into an embedding model via training on MMEB. Unlike previous models such as CLIP and BLIP, VLM2Vec can process any combination of images and text to generate a fixed-dimensional vector based on task instructions. We build a series of VLM2Vec models on Phi-3.5-V and evaluate them on MMEB's evaluation split. Our results show that \model achieves an absolute average improvement of 10% to 20% over existing multimodal embedding models on both in-distribution and out-of-distribution datasets in MMEB.

comment: Technical Report

☆ CTC-GMM: CTC guided modality matching for fast and accurate streaming speech translation

Rui Zhao, Jinyu Li, Ruchao Fan, Matt Post

Models for streaming speech translation (ST) can achieve high accuracy and low latency if they're developed with vast amounts of paired audio in the source language and written text in the target language. Yet, these text labels for the target language are often pseudo labels due to the prohibitive cost of manual ST data labeling. In this paper, we introduce a methodology named Connectionist Temporal Classification guided modality matching (CTC-GMM) that enhances the streaming ST model by leveraging extensive machine translation (MT) text data. This technique employs CTC to compress the speech sequence into a compact embedding sequence that matches the corresponding text sequence, allowing us to utilize matched {source-target} language text pairs from the MT corpora to refine the streaming ST model further. Our evaluations with FLEURS and CoVoST2 show that the CTC-GMM approach can increase translation accuracy relatively by 13.9% and 6.4% respectively, while also boosting decoding speed by 59.7% on GPU.

comment: Accepted by IEEE Spoken Language Technology Workshop (SLT 2024)

☆ SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Masks

Fenia Christopoulou, Ronald Cardenas, Gerasimos Lampouras, Haitham Bou-Ammar, Jun Wang

comment: 20 papges, 9 figures, 5 tables. Under Review

☆ Investigating large language models for their competence in extracting grammatically sound sentences from transcribed noisy utterances CoNLL 2024

Alina Wróblewska

Selectively processing noisy utterances while effectively disregarding speech-specific elements poses no considerable challenge for humans, as they exhibit remarkable cognitive abilities to separate semantically significant content from speech-specific noise (i.e. filled pauses, disfluencies, and restarts). These abilities may be driven by mechanisms based on acquired grammatical rules that compose abstract syntactic-semantic structures within utterances. Segments without syntactic and semantic significance are consistently disregarded in these structures. The structures, in tandem with lexis, likely underpin language comprehension and thus facilitate effective communication. In our study, grounded in linguistically motivated experiments, we investigate whether large language models (LLMs) can effectively perform analogical speech comprehension tasks. In particular, we examine the ability of LLMs to extract well-structured utterances from transcriptions of noisy dialogues. We conduct two evaluation experiments in the Polish language scenario, using a~dataset presumably unfamiliar to LLMs to mitigate the risk of data contamination. Our results show that not all extracted utterances are correctly structured, indicating that either LLMs do not fully acquire syntactic-semantic rules or they acquire them but cannot apply them effectively. We conclude that the ability of LLMs to comprehend noisy utterances is still relatively superficial compared to human proficiency in processing them.

comment: Accepted at CoNLL 2024

☆ Explanation sensitivity to the randomness of large language models: the case of journalistic text classification

Jeremie Bogaert, Marie-Catherine de Marneffe, Antonin Descampe, Louis Escouflaire, Cedrick Fairon, Francois-Xavier Standaert

Large language models (LLMs) perform very well in several natural language processing tasks but raise explainability challenges. In this paper, we examine the effect of random elements in the training of LLMs on the explainability of their predictions. We do so on a task of opinionated journalistic text classification in French. Using a fine-tuned CamemBERT model and an explanation method based on relevance propagation, we find that training with different random seeds produces models with similar accuracy but variable explanations. We therefore claim that characterizing the explanations' statistical distribution is needed for the explainability of LLMs. We then explore a simpler model based on textual features which offers stable explanations but is less accurate. Hence, this simpler model corresponds to a different tradeoff between accuracy and explainability. We show that it can be improved by inserting features derived from CamemBERT's explanations. We finally discuss new research directions suggested by our results, in particular regarding the origin of the sensitivity observed in the training randomness.

comment: This paper is a faithful translation of a paper which was peer-reviewed and published in the French journal Traitement Automatique des Langues, n. 64

☆ ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery

Ziru Chen, Shijie Chen, Yuting Ning, Qianheng Zhang, Boshi Wang, Botao Yu, Yifei Li, Zeyi Liao, Chen Wei, Zitong Lu, Vishal Dey, Mingyi Xue, Frazier N. Baker, Benjamin Burns, Daniel Adu-Ampratwum, Xuhui Huang, Xia Ning, Song Gao, Yu Su, Huan Sun

The advancements of language language models (LLMs) have piqued growing interest in developing LLM-based language agents to automate scientific discovery end-to-end, which has sparked both excitement and skepticism about the true capabilities of such agents. In this work, we argue that for an agent to fully automate scientific discovery, it must be able to complete all essential tasks in the workflow. Thus, we call for rigorous assessment of agents on individual tasks in a scientific workflow before making bold claims on end-to-end automation. To this end, we present ScienceAgentBench, a new benchmark for evaluating language agents for data-driven scientific discovery. To ensure the scientific authenticity and real-world relevance of our benchmark, we extract 102 tasks from 44 peer-reviewed publications in four disciplines and engage nine subject matter experts to validate them. We unify the target output for every task to a self-contained Python program file and employ an array of evaluation metrics to examine the generated programs, execution results, and costs. Each task goes through multiple rounds of manual validation by annotators and subject matter experts to ensure its annotation quality and scientific plausibility. We also propose two effective strategies to mitigate data contamination concerns. Using our benchmark, we evaluate five open-weight and proprietary LLMs, each with three frameworks: direct prompting, OpenHands, and self-debug. Given three attempts for each task, the best-performing agent can only solve 32.4% of the tasks independently and 34.3% with expert-provided knowledge. These results underscore the limited capacities of current language agents in generating code for data-driven discovery, let alone end-to-end automation for scientific research.

comment: 55 pages

☆ ZEBRA: Zero-Shot Example-Based Retrieval Augmentation for Commonsense Question Answering EMNLP 2024

Francesco Maria Molfese, Simone Conia, Riccardo Orlando, Roberto Navigli

Current Large Language Models (LLMs) have shown strong reasoning capabilities in commonsense question answering benchmarks, but the process underlying their success remains largely opaque. As a consequence, recent approaches have equipped LLMs with mechanisms for knowledge retrieval, reasoning and introspection, not only to improve their capabilities but also to enhance the interpretability of their outputs. However, these methods require additional training, hand-crafted templates or human-written explanations. To address these issues, we introduce ZEBRA, a zero-shot question answering framework that combines retrieval, case-based reasoning and introspection and dispenses with the need for additional training of the LLM. Given an input question, ZEBRA retrieves relevant question-knowledge pairs from a knowledge base and generates new knowledge by reasoning over the relationships in these pairs. This generated knowledge is then used to answer the input question, improving the model's performance and interpretability. We evaluate our approach across 8 well-established commonsense reasoning benchmarks, demonstrating that ZEBRA consistently outperforms strong LLMs and previous knowledge integration approaches, achieving an average accuracy improvement of up to 4.5 points.

comment: Accepted at EMNLP 2024 Main Conference

☆ TidalDecode: Fast and Accurate LLM Decoding with Position Persistent Sparse Attention

Lijie Yang, Zhihao Zhang, Zhuofu Chen, Zikun Li, Zhihao Jia

Large language models (LLMs) have driven significant advancements across diverse NLP tasks, with long-context models gaining prominence for handling extended inputs. However, the expanding key-value (KV) cache size required by Transformer architectures intensifies the memory constraints, particularly during the decoding phase, creating a significant bottleneck. Existing sparse attention mechanisms designed to address this bottleneck have two limitations: (1) they often fail to reliably identify the most relevant tokens for attention, and (2) they overlook the spatial coherence of token selection across consecutive Transformer layers, which can lead to performance degradation and substantial overhead in token selection. This paper introduces TidalDecode, a simple yet effective algorithm and system for fast and accurate LLM decoding through position persistent sparse attention. TidalDecode leverages the spatial coherence of tokens selected by existing sparse attention methods and introduces a few token selection layers that perform full attention to identify the tokens with the highest attention scores, while all other layers perform sparse attention with the pre-selected tokens. This design enables TidalDecode to substantially reduce the overhead of token selection for sparse attention without sacrificing the quality of the generated results. Evaluation on a diverse set of LLMs and tasks shows that TidalDecode closely matches the generative performance of full attention methods while reducing the LLM decoding latency by up to 2.1x.

☆ Initialization of Large Language Models via Reparameterization to Mitigate Loss Spikes EMNLP2024

Kosuke Nishida, Kyosuke Nishida, Kuniko Saito

Loss spikes, a phenomenon in which the loss value diverges suddenly, is a fundamental issue in the pre-training of large language models. This paper supposes that the non-uniformity of the norm of the parameters is one of the causes of loss spikes. Here, in training of neural networks, the scale of the gradients is required to be kept constant throughout the layers to avoid the vanishing and exploding gradients problem. However, to meet these requirements in the Transformer model, the norm of the model parameters must be non-uniform, and thus, parameters whose norm is smaller are more sensitive to the parameter update. To address this issue, we propose a novel technique, weight scaling as reparameterization (WeSaR). WeSaR introduces a gate parameter per parameter matrix and adjusts it to the value satisfying the requirements. Because of the gate parameter, WeSaR sets the norm of the original parameters uniformly, which results in stable training. Experimental results with the Transformer decoders consisting of 130 million, 1.3 billion, and 13 billion parameters showed that WeSaR stabilizes and accelerates training and that it outperformed compared methods including popular initialization methods.

comment: EMNLP2024 accepted

☆ A test suite of prompt injection attacks for LLM-based machine translation

Antonio Valerio Miceli-Barone, Zhifan Sun

LLM-based NLP systems typically work by embedding their input data into prompt templates which contain instructions and/or in-context examples, creating queries which are submitted to a LLM, and then parsing the LLM response in order to generate the system outputs. Prompt Injection Attacks (PIAs) are a type of subversion of these systems where a malicious user crafts special inputs which interfere with the prompt templates, causing the LLM to respond in ways unintended by the system designer. Recently, Sun and Miceli-Barone proposed a class of PIAs against LLM-based machine translation. Specifically, the task is to translate questions from the TruthfulQA test suite, where an adversarial prompt is prepended to the questions, instructing the system to ignore the translation instruction and answer the questions instead. In this test suite, we extend this approach to all the language pairs of the WMT 2024 General Machine Translation task. Moreover, we include additional attack formats in addition to the one originally studied.

☆ Named Clinical Entity Recognition Benchmark

Wadood M Abdul, Marco AF Pimentel, Muhammad Umar Salman, Tathagata Raha, Clément Christophe, Praveen K Kanithi, Nasir Hayat, Ronnie Rajan, Shadab Khan

This technical report introduces a Named Clinical Entity Recognition Benchmark for evaluating language models in healthcare, addressing the crucial natural language processing (NLP) task of extracting structured information from clinical narratives to support applications like automated coding, clinical trial cohort identification, and clinical decision support. The leaderboard provides a standardized platform for assessing diverse language models, including encoder and decoder architectures, on their ability to identify and classify clinical entities across multiple medical domains. A curated collection of openly available clinical datasets is utilized, encompassing entities such as diseases, symptoms, medications, procedures, and laboratory measurements. Importantly, these entities are standardized according to the Observational Medical Outcomes Partnership (OMOP) Common Data Model, ensuring consistency and interoperability across different healthcare systems and datasets, and a comprehensive evaluation of model performance. Performance of models is primarily assessed using the F1-score, and it is complemented by various assessment modes to provide comprehensive insights into model performance. The report also includes a brief analysis of models evaluated to date, highlighting observed trends and limitations. By establishing this benchmarking framework, the leaderboard aims to promote transparency, facilitate comparative analyses, and drive innovation in clinical entity recognition tasks, addressing the need for robust evaluation methods in healthcare NLP.

comment: Technical Report

☆ Can LLMs plan paths with extra hints from solvers?

Erik Wu, Sayan Mitra

Large Language Models (LLMs) have shown remarkable capabilities in natural language processing, mathematical problem solving, and tasks related to program synthesis. However, their effectiveness in long-term planning and higher-order reasoning has been noted to be limited and fragile. This paper explores an approach for enhancing LLM performance in solving a classical robotic planning task by integrating solver-generated feedback. We explore four different strategies for providing feedback, including visual feedback, we utilize fine-tuning, and we evaluate the performance of three different LLMs across a 10 standard and 100 more randomly generated planning problems. Our results suggest that the solver-generated feedback improves the LLM's ability to solve the moderately difficult problems, but the harder problems still remain out of reach. The study provides detailed analysis of the effects of the different hinting strategies and the different planning tendencies of the evaluated LLMs.

☆ DEPT: Decoupled Embeddings for Pre-training Language Models

Alex Iacob, Lorenzo Sani, Meghdad Kurmanji, William F. Shen, Xinchi Qiu, Dongqi Cai, Yan Gao, Nicholas D. Lane

☆ On the Biased Assessment of Expert Finding Systems RecSys

Jens-Joris Decorte, Jeroen Van Hautte, Chris Develder, Thomas Demeester

In large organisations, identifying experts on a given topic is crucial in leveraging the internal knowledge spread across teams and departments. So-called enterprise expert retrieval systems automatically discover and structure employees' expertise based on the vast amount of heterogeneous data available about them and the work they perform. Evaluating these systems requires comprehensive ground truth expert annotations, which are hard to obtain. Therefore, the annotation process typically relies on automated recommendations of knowledge areas to validate. This case study provides an analysis of how these recommendations can impact the evaluation of expert finding systems. We demonstrate on a popular benchmark that system-validated annotations lead to overestimated performance of traditional term-based retrieval models and even invalidate comparisons with more recent neural methods. We also augment knowledge areas with synonyms to uncover a strong bias towards literal mentions of their constituent words. Finally, we propose constraints to the annotation process to prevent these biased evaluations, and show that this still allows annotation suggestions of high utility. These findings should inform benchmark creation or selection for expert finding, to guarantee meaningful comparison of methods.

comment: Accepted to the 4th Workshop on Recommender Systems for Human Resources (RecSys in HR 2024) as part of RecSys 2024

☆ SkillMatch: Evaluating Self-supervised Learning of Skill Relatedness ECML-PKDD 2024

Jens-Joris Decorte, Jeroen Van Hautte, Thomas Demeester, Chris Develder

Accurately modeling the relationships between skills is a crucial part of human resources processes such as recruitment and employee development. Yet, no benchmarks exist to evaluate such methods directly. We construct and release SkillMatch, a benchmark for the task of skill relatedness, based on expert knowledge mining from millions of job ads. Additionally, we propose a scalable self-supervised learning technique to adapt a Sentence-BERT model based on skill co-occurrence in job ads. This new method greatly surpasses traditional models for skill relatedness as measured on SkillMatch. By releasing SkillMatch publicly, we aim to contribute a foundation for research towards increased accuracy and transparency of skill-based recommendation systems.

comment: Accepted to the International workshop on AI for Human Resources and Public Employment Services (AI4HR&PES) as part of ECML-PKDD 2024

☆ On the Rigour of Scientific Writing: Criteria, Analysis, and Insights EMNLP 2024

Joseph James, Chenghao Xiao, Yucheng Li, Chenghua Lin

Rigour is crucial for scientific research as it ensures the reproducibility and validity of results and findings. Despite its importance, little work exists on modelling rigour computationally, and there is a lack of analysis on whether these criteria can effectively signal or measure the rigour of scientific papers in practice. In this paper, we introduce a bottom-up, data-driven framework to automatically identify and define rigour criteria and assess their relevance in scientific writing. Our framework includes rigour keyword extraction, detailed rigour definition generation, and salient criteria identification. Furthermore, our framework is domain-agnostic and can be tailored to the evaluation of scientific rigour for different areas, accommodating the distinct salient criteria across fields. We conducted comprehensive experiments based on datasets collected from two high impact venues for Machine Learning and NLP (i.e., ICLR and ACL) to demonstrate the effectiveness of our framework in modelling rigour. In addition, we analyse linguistic patterns of rigour, revealing that framing certainty is crucial for enhancing the perception of scientific rigour, while suggestion certainty and probability uncertainty diminish it.

comment: Accepted Findings at EMNLP 2024

☆ Activation Scaling for Steering and Interpreting Language Models EMNLP 2024

Niklas Stoehr, Kevin Du, Vésteinn Snæbjarnarson, Robert West, Ryan Cotterell, Aaron Schein

Given the prompt "Rome is in", can we steer a language model to flip its prediction of an incorrect token "France" to a correct token "Italy" by only multiplying a few relevant activation vectors with scalars? We argue that successfully intervening on a model is a prerequisite for interpreting its internal workings. Concretely, we establish a three-term objective: a successful intervention should flip the correct with the wrong token and vice versa (effectiveness), and leave other tokens unaffected (faithfulness), all while being sparse (minimality). Using gradient-based optimization, this objective lets us learn (and later evaluate) a specific kind of efficient and interpretable intervention: activation scaling only modifies the signed magnitude of activation vectors to strengthen, weaken, or reverse the steering directions already encoded in the model. On synthetic tasks, this intervention performs comparably with steering vectors in terms of effectiveness and faithfulness, but is much more minimal allowing us to pinpoint interpretable model components. We evaluate activation scaling from different angles, compare performance on different datasets, and make activation scalars a learnable function of the activation vectors themselves to generalize to varying-length prompts.

comment: Findings of the Association for Computational Linguistics: EMNLP 2024

☆ Intent Classification for Bank Chatbots through LLM Fine-Tuning

Bibiána Lajčinová, Patrik Valábek, Michal Spišiak

This study evaluates the application of large language models (LLMs) for intent classification within a chatbot with predetermined responses designed for banking industry websites. Specifically, the research examines the effectiveness of fine-tuning SlovakBERT compared to employing multilingual generative models, such as Llama 8b instruct and Gemma 7b instruct, in both their pre-trained and fine-tuned versions. The findings indicate that SlovakBERT outperforms the other models in terms of in-scope accuracy and out-of-scope false positive rate, establishing it as the benchmark for this application.

comment: 7 pages, no figures

☆ Leveraging Grammar Induction for Language Understanding and Generation EMNLP 2024

Jushi Kai, Shengyuan Hou, Yusheng Huang, Zhouhan Lin

Grammar induction has made significant progress in recent years. However, it is not clear how the application of induced grammar could enhance practical performance in downstream tasks. In this work, we introduce an unsupervised grammar induction method for language understanding and generation. We construct a grammar parser to induce constituency structures and dependency relations, which is simultaneously trained on downstream tasks without additional syntax annotations. The induced grammar features are subsequently incorporated into Transformer as a syntactic mask to guide self-attention. We evaluate and apply our method to multiple machine translation tasks and natural language understanding tasks. Our method demonstrates superior performance compared to the original Transformer and other models enhanced with external parsers. Experimental results indicate that our method is effective in both from-scratch and pre-trained scenarios. Additionally, our research highlights the contribution of explicitly modeling the grammatical structure of texts to neural network models.

comment: EMNLP 2024 Findings

☆ Rationale-Aware Answer Verification by Pairwise Self-Evaluation EMNLP 2024

Akira Kawabata, Saku Sugawara

Answer verification identifies correct solutions among candidates generated by large language models (LLMs). Current approaches typically train verifier models by labeling solutions as correct or incorrect based solely on whether the final answer matches the gold answer. However, this approach neglects any flawed rationale in the solution yielding the correct answer, undermining the verifier's ability to distinguish between sound and flawed rationales. We empirically show that in StrategyQA, only 19% of LLM-generated solutions with correct answers have valid rationales, thus leading to an unreliable verifier. Furthermore, we demonstrate that training a verifier on valid rationales significantly improves its ability to distinguish valid and flawed rationale. To make a better verifier without extra human supervision, we introduce REPS (Rationale Enhancement through Pairwise Selection), a method for selecting valid rationales from candidates by iteratively applying pairwise self-evaluation using the same LLM that generates the solutions. Verifiers trained on solutions selected by REPS outperform those trained using conventional training methods on three reasoning benchmarks (ARC-Challenge, DROP, and StrategyQA). Our results suggest that training reliable verifiers requires ensuring the validity of rationales in addition to the correctness of the final answers, which would be critical for models assisting humans in solving complex reasoning tasks.

comment: EMNLP 2024

☆ As Simple as Fine-tuning: LLM Alignment via Bidirectional Negative Feedback Loss

Xin Mao, Feng-Lin Li, Huimin Xu, Wei Zhang, Wang Chen, Anh Tuan Luu

Direct Preference Optimization (DPO) has emerged as a more computationally efficient alternative to Reinforcement Learning from Human Feedback (RLHF) with Proximal Policy Optimization (PPO), eliminating the need for reward models and online sampling. Despite these benefits, DPO and its variants remain sensitive to hyper-parameters and prone to instability, particularly on mathematical datasets. We argue that these issues arise from the unidirectional likelihood-derivative negative feedback inherent in the log-likelihood loss function. To address this, we propose a novel LLM alignment loss that establishes a stable Bidirectional Negative Feedback (BNF) during optimization. Our proposed BNF loss eliminates the need for pairwise contrastive losses and does not require any extra tunable hyper-parameters or pairwise preference data, streamlining the alignment pipeline to be as simple as supervised fine-tuning. We conduct extensive experiments across two challenging QA benchmarks and four reasoning benchmarks. The experimental results show that BNF achieves comparable performance to the best methods on QA benchmarks, while its performance decrease on the four reasoning benchmarks is significantly lower compared to the best methods, thus striking a better balance between value alignment and reasoning ability. In addition, we further validate the performance of BNF on non-pairwise datasets, and conduct in-depth analysis of log-likelihood and logit shifts across different preference optimization methods.

comment: 20 pages, 9 figures

☆ MINER: Mining the Underlying Pattern of Modality-Specific Neurons in Multimodal Large Language Models

Kaichen Huang, Jiahao Huo, Yibo Yan, Kun Wang, Yutao Yue, Xuming Hu

In recent years, multimodal large language models (MLLMs) have significantly advanced, integrating more modalities into diverse applications. However, the lack of explainability remains a major barrier to their use in scenarios requiring decision transparency. Current neuron-level explanation paradigms mainly focus on knowledge localization or language- and domain-specific analyses, leaving the exploration of multimodality largely unaddressed. To tackle these challenges, we propose MINER, a transferable framework for mining modality-specific neurons (MSNs) in MLLMs, which comprises four stages: (1) modality separation, (2) importance score calculation, (3) importance score aggregation, (4) modality-specific neuron selection. Extensive experiments across six benchmarks and two representative MLLMs show that (I) deactivating ONLY 2% of MSNs significantly reduces MLLMs performance (0.56 to 0.24 for Qwen2-VL, 0.69 to 0.31 for Qwen2-Audio), (II) different modalities mainly converge in the lower layers, (III) MSNs influence how key information from various modalities converges to the last token, (IV) two intriguing phenomena worth further investigation, i.e., semantic probing and semantic telomeres. The source code is available at this URL.

☆ LPZero: Language Model Zero-cost Proxy Search from Zero

Peijie Dong, Lujun Li, Xiang Liu, Zhenheng Tang, Xuebo Liu, Qiang Wang, Xiaowen Chu

In spite of the outstanding performance, Neural Architecture Search (NAS) is criticized for massive computation. Recently, Zero-shot NAS has emerged as a promising approach by exploiting Zero-cost (ZC) proxies, which markedly reduce computational demands. Despite this, existing ZC proxies heavily rely on expert knowledge and incur significant trial-and-error costs. Particularly in NLP tasks, most existing ZC proxies fail to surpass the performance of the naive baseline. To address these challenges, we introduce a novel framework, \textbf{LPZero}, which is the first to automatically design ZC proxies for various tasks, achieving higher ranking consistency than human-designed proxies. Specifically, we model the ZC proxy as a symbolic equation and incorporate a unified proxy search space that encompasses existing ZC proxies, which are composed of a predefined set of mathematical symbols. To heuristically search for the best ZC proxy, LPZero incorporates genetic programming to find the optimal symbolic composition. We propose a \textit{Rule-based Pruning Strategy (RPS),} which preemptively eliminates unpromising proxies, thereby mitigating the risk of proxy degradation. Extensive experiments on FlexiBERT, GPT-2, and LLaMA-7B demonstrate LPZero's superior ranking ability and performance on downstream tasks compared to current approaches.

comment: 8 pages, 7 figures, 10 appendix

☆ DAPE V2: Process Attention Score as Feature Map for Length Extrapolation

Chuanyang Zheng, Yihang Gao, Han Shi, Jing Xiong, Jiankai Sun, Jingyao Li, Minbin Huang, Xiaozhe Ren, Michael Ng, Xin Jiang, Zhenguo Li, Yu Li

The attention mechanism is a fundamental component of the Transformer model, contributing to interactions among distinct tokens, in contrast to earlier feed-forward neural networks. In general, the attention scores are determined simply by the key-query products. However, this work's occasional trial (combining DAPE and NoPE) of including additional MLPs on attention scores without position encoding indicates that the classical key-query multiplication may limit the performance of Transformers. In this work, we conceptualize attention as a feature map and apply the convolution operator (for neighboring attention scores across different heads) to mimic the processing methods in computer vision. Specifically, the main contribution of this paper is identifying and interpreting the Transformer length extrapolation problem as a result of the limited expressiveness of the naive query and key dot product, and we successfully translate the length extrapolation issue into a well-understood feature map processing problem. The novel insight, which can be adapted to various attention-related models, reveals that the current Transformer architecture has the potential for further evolution. Extensive experiments demonstrate that treating attention as a feature map and applying convolution as a processing method significantly enhances Transformer performance.

comment: Tech Report. arXiv admin note: text overlap with arXiv:2405.14722

☆ Representing the Under-Represented: Cultural and Core Capability Benchmarks for Developing Thai Large Language Models

Dahyun Kim, Sukyung Lee, Yungi Kim, Attapol Rutherford, Chanjun Park

☆ GARLIC: LLM-Guided Dynamic Progress Control with Hierarchical Weighted Graph for Long Document QA

Xinyu Wang, Yanzheng Xiang, Lin Gui, Yulan He

In the past, Retrieval-Augmented Generation (RAG) methods split text into chunks to enable language models to handle long documents. Recent tree-based RAG methods are able to retrieve detailed information while preserving global context. However, with the advent of more powerful LLMs, such as Llama 3.1, which offer better comprehension and support for longer inputs, we found that even recent tree-based RAG methods perform worse than directly feeding the entire document into Llama 3.1, although RAG methods still hold an advantage in reducing computational costs. In this paper, we propose a new retrieval method, called LLM-Guided Dynamic Progress Control with Hierarchical Weighted Graph (GARLIC), which outperforms previous state-of-the-art baselines, including Llama 3.1, while retaining the computational efficiency of RAG methods. Our method introduces several improvements: (1) Rather than using a tree structure, we construct a Hierarchical Weighted Directed Acyclic Graph with many-to-many summarization, where the graph edges are derived from attention mechanisms, and each node focuses on a single event or very few events. (2) We introduce a novel retrieval method that leverages the attention weights of LLMs rather than dense embedding similarity. Our method allows for searching the graph along multiple paths and can terminate at any depth. (3) We use the LLM to control the retrieval process, enabling it to dynamically adjust the amount and depth of information retrieved for different queries. Experimental results show that our method outperforms previous state-of-the-art baselines, including Llama 3.1, on two single-document and two multi-document QA datasets, while maintaining similar computational complexity to traditional RAG methods.

☆ Formality is Favored: Unraveling the Learning Preferences of Large Language Models on Data with Conflicting Knowledge EMNLP 2024

Jiahuan Li, Yiqing Cao, Shujian Huang, Jiajun Chen

Having been trained on massive pretraining data, large language models have shown excellent performance on many knowledge-intensive tasks. However, pretraining data tends to contain misleading and even conflicting information, and it is intriguing to understand how LLMs handle these noisy data during training. In this study, we systematically analyze LLMs' learning preferences for data with conflicting knowledge. We find that pretrained LLMs establish learning preferences similar to humans, i.e., preferences towards formal texts and texts with fewer spelling errors, resulting in faster learning and more favorable treatment of knowledge in data with such features when facing conflicts. This finding is generalizable across models and languages and is more evident in larger models. An in-depth analysis reveals that LLMs tend to trust data with features that signify consistency with the majority of data, and it is possible to instill new preferences and erase old ones by manipulating the degree of consistency with the majority data.

comment: accepted by EMNLP 2024, main conference

☆ ImProver: Agent-Based Automated Proof Optimization

Riyaz Ahuja, Jeremy Avigad, Prasad Tetali, Sean Welleck

Large language models (LLMs) have been used to generate formal proofs of mathematical theorems in proofs assistants such as Lean. However, we often want to optimize a formal proof with respect to various criteria, depending on its downstream use. For example, we may want a proof to adhere to a certain style, or to be readable, concise, or modularly structured. Having suitably optimized proofs is also important for learning tasks, especially since human-written proofs may not optimal for that purpose. To this end, we study a new problem of automated proof optimization: rewriting a proof so that it is correct and optimizes for an arbitrary criterion, such as length or readability. As a first method for automated proof optimization, we present ImProver, a large-language-model agent that rewrites proofs to optimize arbitrary user-defined metrics in Lean. We find that naively applying LLMs to proof optimization falls short, and we incorporate various improvements into ImProver, such as the use of symbolic Lean context in a novel Chain-of-States technique, as well as error-correction and retrieval. We test ImProver on rewriting real-world undergraduate, competition, and research-level mathematics theorems, finding that ImProver is capable of rewriting proofs so that they are substantially shorter, more modular, and more readable.

comment: 19 pages, 21 figures

☆ Document-level Causal Relation Extraction with Knowledge-guided Binary Question Answering EMNLP 2024

Zimu Wang, Lei Xia, Wei Wang, Xinya Du

As an essential task in information extraction (IE), Event-Event Causal Relation Extraction (ECRE) aims to identify and classify the causal relationships between event mentions in natural language texts. However, existing research on ECRE has highlighted two critical challenges, including the lack of document-level modeling and causal hallucinations. In this paper, we propose a Knowledge-guided binary Question Answering (KnowQA) method with event structures for ECRE, consisting of two stages: Event Structure Construction and Binary Question Answering. We conduct extensive experiments under both zero-shot and fine-tuning settings with large language models (LLMs) on the MECI and MAVEN-ERE datasets. Experimental results demonstrate the usefulness of event structures on document-level ECRE and the effectiveness of KnowQA by achieving state-of-the-art on the MECI dataset. We observe not only the effectiveness but also the high generalizability and low inconsistency of our method, particularly when with complete event structures after fine-tuning the models.

comment: Accepted at Findings of EMNLP 2024. Camera-ready version

☆ Intriguing Properties of Large Language and Vision Models

Young-Jun Lee, Byungsoo Ko, Han-Gyu Kim, Yechan Hwang, Ho-Jin Choi

Recently, large language and vision models (LLVMs) have received significant attention and development efforts due to their remarkable generalization performance across a wide range of tasks requiring perception and cognitive abilities. A key factor behind their success is their simple architecture, which consists of a vision encoder, a projector, and a large language model (LLM). Despite their achievements in advanced reasoning tasks, their performance on fundamental perception-related tasks (e.g., MMVP) remains surprisingly low. This discrepancy raises the question of how LLVMs truly perceive images and exploit the advantages of the vision encoder. To address this, we systematically investigate this question regarding several aspects: permutation invariance, robustness, math reasoning, alignment preserving and importance, by evaluating the most common LLVM's families (i.e., LLaVA) across 10 evaluation benchmarks. Our extensive experiments reveal several intriguing properties of current LLVMs: (1) they internally process the image in a global manner, even when the order of visual patch sequences is randomly permuted; (2) they are sometimes able to solve math problems without fully perceiving detailed numerical information; (3) the cross-modal alignment is overfitted to complex reasoning tasks, thereby, causing them to lose some of the original perceptual capabilities of their vision encoder; (4) the representation space in the lower layers (<25%) plays a crucial role in determining performance and enhancing visual understanding. Lastly, based on the above observations, we suggest potential future directions for building better LLVMs and constructing more challenging evaluation benchmarks.

comment: Code is available in https://github.com/passing2961/IP-LLVM

☆ TableRAG: Million-Token Table Understanding with Language Models NeurIPS 2024

Si-An Chen, Lesly Miculicich, Julian Martin Eisenschlos, Zifeng Wang, Zilong Wang, Yanfei Chen, Yasuhisa Fujii, Hsuan-Tien Lin, Chen-Yu Lee, Tomas Pfister

Recent advancements in language models (LMs) have notably enhanced their ability to reason with tabular data, primarily through program-aided mechanisms that manipulate and analyze tables. However, these methods often require the entire table as input, leading to scalability challenges due to the positional bias or context length constraints. In response to these challenges, we introduce TableRAG, a Retrieval-Augmented Generation (RAG) framework specifically designed for LM-based table understanding. TableRAG leverages query expansion combined with schema and cell retrieval to pinpoint crucial information before providing it to the LMs. This enables more efficient data encoding and precise retrieval, significantly reducing prompt lengths and mitigating information loss. We have developed two new million-token benchmarks from the Arcade and BIRD-SQL datasets to thoroughly evaluate TableRAG's effectiveness at scale. Our results demonstrate that TableRAG's retrieval design achieves the highest retrieval quality, leading to the new state-of-the-art performance on large-scale table understanding.

comment: Accepted to NeurIPS 2024

☆ TLDR: Token-Level Detective Reward Model for Large Vision Language Models

Deqing Fu, Tong Xiao, Rui Wang, Wang Zhu, Pengchuan Zhang, Guan Pang, Robin Jia, Lawrence Chen

Although reward models have been successful in improving multimodal large language models, the reward models themselves remain brutal and contain minimal information. Notably, existing reward models only mimic human annotations by assigning only one binary feedback to any text, no matter how long the text is. In the realm of multimodal language models, where models are required to process both images and texts, a naive reward model may learn implicit biases toward texts and become less grounded in images. In this paper, we propose a $\textbf{T}$oken-$\textbf{L}$evel $\textbf{D}$etective $\textbf{R}$eward Model ($\textbf{TLDR}$) to provide fine-grained annotations to each text token. We first introduce a perturbation-based method to generate synthetic hard negatives and their token-level labels to train TLDR models. Then we show the rich usefulness of TLDR models both in assisting off-the-shelf models to self-correct their generations, and in serving as a hallucination evaluation tool. Finally, we show that TLDR models can significantly speed up human annotation by 3 times to acquire a broader range of high-quality vision language data.

comment: Work done at Meta

☆ Efficient transformer with reinforced position embedding for language models

Yen-Che Hsiao, Abhishek Dutta

In this paper, we propose an efficient transformer architecture that uses reinforced positional embedding to obtain superior performance with half the number of encoder decoder layers. We demonstrate that concatenating positional encoding with trainable token embeddings, normalizing columns in the token embedding matrix, and using the normalized token embedding matrix as the value of the attention layer improve the training and validation loss and the training time in an encoder-decoder Transformer model for a Portuguese-English translation task with 10 epochs or 12 hours of training across 10 trials. Our method, with roughly a threefold parameter reduction compared to the baseline model, yields a mean training loss of 1.21, a mean validation loss of 1.51, and an average training time of 1352.27 seconds per epoch, surpassing the baseline model with the same embedding dimension that employs addition of positional encoding and token embeddings, which achieves a mean training loss of 1.96, a validation loss of 2.18, and an average training time of 4297.79 seconds per epoch. Additionally, we evaluated our proposed architecture and the baseline across 14 diverse translation datasets from TensorFlow. The results indicate that our method consistently achieves lower or comparable training and validation losses, suggesting enhanced learning efficiency.

☆ Forgetting Curve: A Reliable Method for Evaluating Memorization Capability for Long-context Models

Xinyu Liu, Runsong Zhao, Pengcheng Huang, Chunyang Xiao, Bei Li, Jingang Wang, Tong Xiao, Jingbo Zhu

Numerous recent works target to extend effective context length for language models and various methods, tasks and benchmarks exist to measure model's effective memorization length. However, through thorough investigations, we find limitations for currently existing evaluations on model's memorization capability. We provide an extensive survey for limitations in this work and propose a new method called forgetting curve to measure the memorization capability of long-context models. We show that forgetting curve has the advantage of being robust to the tested corpus and the experimental settings, of not relying on prompts and can be applied to any model size. We apply our forgetting curve to a large variety of models involving both transformer and RNN/SSM based architectures. Our measurement provides empirical evidence for the effectiveness of transformer extension techniques while raises questions for the effective length of RNN/SSM based models. We also examine the difference between our measurement and existing benchmarks as well as popular metrics for various models. Our code and results can be found at https://github.com/1azybug/ForgettingCurve.

☆ $\textbf{Only-IF}$:Revealing the Decisive Effect of Instruction Diversity on Generalization

Dylan Zhang, Justin Wang, Francois Charton

Understanding and accurately following instructions is critical for large language models (LLMs) to be effective across diverse tasks. In this work, we rigorously examine the key factors that enable models to generalize to unseen instructions, providing insights to guide the collection of data for instruction-tuning. Through controlled experiments, inspired by the Turing-complete Markov algorithm, we demonstrate that such generalization $\textbf{only emerges}$ when training data is diversified enough across semantic domains. Our findings also reveal that merely diversifying within limited domains fails to ensure robust generalization. In contrast, cross-domain data diversification, even under constrained data budgets, significantly enhances a model's adaptability. We further extend our analysis to real-world scenarios, including fine-tuning of $\textit{$\textbf{specialist}$}$ and $\textit{$\textbf{generalist}$}$ models. In both cases, we demonstrate that 1) better performance can be achieved by increasing the diversity of an established dataset while keeping the data size constant, and 2) when scaling up the data, diversifying the semantics of instructions is more effective than simply increasing the quantity of similar data. Our research provides important insights for dataset collation, particularly when optimizing model performance by expanding training data for both specialist and generalist scenarios. We show that careful consideration of data diversification is key: training specialist models with data extending beyond their core domain leads to significant performance improvements, while generalist models benefit from diverse data mixtures that enhance their overall instruction-following capabilities across a wide range of applications. Our results highlight the critical role of strategic diversification and offer clear guidelines for improving data quality.

☆ Rule-based Data Selection for Large Language Models

Xiaomin Li, Mingye Gao, Zhiwei Zhang, Chang Yue, Hong Hu

The quality of training data significantly impacts the performance of large language models (LLMs). There are increasing studies using LLMs to rate and select data based on several human-crafted metrics (rules). However, these conventional rule-based approaches often depend too heavily on human heuristics, lack effective metrics for assessing rules, and exhibit limited adaptability to new tasks. In our study, we introduce an innovative rule-based framework that utilizes the orthogonality of score vectors associated with rules as a novel metric for rule evaluations. Our approach includes an automated pipeline that first uses LLMs to generate a diverse set of rules, encompassing various rating dimensions to evaluate data quality. Then it rates a batch of data based on these rules and uses the determinantal point process (DPP) from random matrix theory to select the most orthogonal score vectors, thereby identifying a set of independent rules. These rules are subsequently used to evaluate all data, selecting samples with the highest average scores for downstream tasks such as LLM training. We verify the effectiveness of our method through two experimental setups: 1) comparisons with ground truth ratings and 2) benchmarking LLMs trained with the chosen data. Our comprehensive experiments cover a range of scenarios, including general pre-training and domain-specific fine-tuning in areas such as IMDB, Medical, Math, and Code. The outcomes demonstrate that our DPP-based rule rating method consistently outperforms other approaches, including rule-free rating, uniform sampling, importance resampling, and QuRating, in terms of both rating precision and model performance.

☆ Learning How Hard to Think: Input-Adaptive Allocation of LM Computation

Mehul Damani, Idan Shenfeld, Andi Peng, Andreea Bobu, Jacob Andreas

Computationally intensive decoding procedures--including search, reranking, and self-critique--can improve the quality of language model (LM) outputs in problems spanning code generation, numerical reasoning, and dialog. Existing work typically applies the same decoding procedure for every input to an LM. But not all inputs require the same amount of computation to process. Can we allocate decoding computation adaptively, using more resources to answer questions whose answers will be harder to compute? We present an approach that predicts the distribution of rewards given an input and computation budget, then allocates additional computation to inputs for which it is predicted to be most useful. We apply this approach in two decoding procedures: first, an adaptive best-of-k procedure that dynamically selects the number of samples to generate as input to a reranker; second, a routing procedure that dynamically responds to a query using a decoding procedure that is expensive but accurate, or one that is cheaper but less capable. Across a suite of programming, mathematics, and dialog tasks, we show that accurate computation-allocation procedures can be learned, and reduce computation by up to 50% at no cost to response quality, or improve quality by up to 10% at a fixed computational budget.

☆ Modeling and Estimation of Vocal Tract and Glottal Source Parameters Using ARMAX-LF Model

Kai Lia, Masato Akagia, Yongwei Lib, Masashi Unokia

Modeling and estimation of the vocal tract and glottal source parameters of vowels from raw speech can be typically done by using the Auto-Regressive with eXogenous input (ARX) model and Liljencrants-Fant (LF) model with an iteration-based estimation approach. However, the all-pole autoregressive model in the modeling of vocal tract filters cannot provide the locations of anti-formants (zeros), which increases the estimation errors in certain classes of speech sounds, such as nasal, fricative, and stop consonants. In this paper, we propose the Auto-Regressive Moving Average eXogenous with LF (ARMAX-LF) model to extend the ARX-LF model to a wider variety of speech sounds, including vowels and nasalized consonants. The LF model represents the glottal source derivative as a parametrized time-domain model, and the ARMAX model represents the vocal tract as a pole-zero filter with an additional exogenous LF excitation as input. To estimate multiple parameters with fewer errors, we first utilize the powerful nonlinear fitting ability of deep neural networks (DNNs) to build a mapping from extracted glottal source derivatives or speech waveforms to corresponding LF parameters. Then, glottal source and vocal tract parameters can be estimated with fewer estimation errors and without any iterations as in the analysis-by-synthesis strategy. Experimental results with synthesized speech using the linear source-filter model, synthesized speech using the physical model, and real speech signals showed that the proposed ARMAX-LF model with a DNN-based estimation method can estimate the parameters of both vowels and nasalized sounds with fewer errors and estimation time.

☆ The LLM Effect: Are Humans Truly Using LLMs, or Are They Being Influenced By Them Instead? EMNLP

Alexander S. Choi, Syeda Sabrina Akter, JP Singh, Antonios Anastasopoulos

Large Language Models (LLMs) have shown capabilities close to human performance in various analytical tasks, leading researchers to use them for time and labor-intensive analyses. However, their capability to handle highly specialized and open-ended tasks in domains like policy studies remains in question. This paper investigates the efficiency and accuracy of LLMs in specialized tasks through a structured user study focusing on Human-LLM partnership. The study, conducted in two stages-Topic Discovery and Topic Assignment-integrates LLMs with expert annotators to observe the impact of LLM suggestions on what is usually human-only analysis. Results indicate that LLM-generated topic lists have significant overlap with human generated topic lists, with minor hiccups in missing document-specific topics. However, LLM suggestions may significantly improve task completion speed, but at the same time introduce anchoring bias, potentially affecting the depth and nuance of the analysis, raising a critical question about the trade-off between increased efficiency and the risk of biased analysis.

comment: Accepted to EMNLP Main 2024. First two authors contributed equally

☆ MathHay: An Automated Benchmark for Long-Context Mathematical Reasoning in LLMs

Lei Wang, Shan Dong, Yuhui Xu, Hanze Dong, Yalu Wang, Amrita Saha, Ee-Peng Lim, Caiming Xiong, Doyen Sahoo

Recent large language models (LLMs) have demonstrated versatile capabilities in long-context scenarios. Although some recent benchmarks have been developed to evaluate the long-context capabilities of LLMs, there is a lack of benchmarks evaluating the mathematical reasoning abilities of LLMs over long contexts, which is crucial for LLMs' application in real-world scenarios. In this paper, we introduce MathHay, an automated benchmark designed to assess the long-context mathematical reasoning capabilities of LLMs. Unlike previous benchmarks like Needle in a Haystack, which focus primarily on information retrieval within long texts, MathHay demands models with both information-seeking and complex mathematical reasoning abilities. We conduct extensive experiments on MathHay to assess the long-context mathematical reasoning abilities of eight top-performing LLMs. Even the best-performing model, Gemini-1.5-Pro-002, still struggles with mathematical reasoning over long contexts, achieving only 51.26% accuracy at 128K tokens. This highlights the significant room for improvement on the MathHay benchmark.

comment: Work-in-Progress

☆ Deeper Insights Without Updates: The Power of In-Context Learning Over Fine-Tuning EMNLP'24

Qingyu Yin, Xuzheng He, Luoao Deng, Chak Tou Leong, Fan Wang, Yanzhao Yan, Xiaoyu Shen, Qiang Zhang

Fine-tuning and in-context learning (ICL) are two prevalent methods in imbuing large language models with task-specific knowledge. It is commonly believed that fine-tuning can surpass ICL given sufficient training samples as it allows the model to adjust its internal parameters based on the data. However, this paper presents a counterintuitive finding: For tasks with implicit patterns, ICL captures these patterns significantly better than fine-tuning. We developed several datasets featuring implicit patterns, such as sequences determining answers through parity or identifying reducible terms in calculations. We then evaluated the models' understanding of these patterns under both fine-tuning and ICL across models ranging from 0.5B to 7B parameters. The results indicate that models employing ICL can quickly grasp deep patterns and significantly improve accuracy. In contrast, fine-tuning, despite utilizing thousands of times more training samples than ICL, achieved only limited improvements. We also proposed circuit shift theory from a mechanistic interpretability's view to explain why ICL wins.

comment: EMNLP'24 Findings

☆ Adversarial Multi-Agent Evaluation of Large Language Models through Iterative Debates

Chaithanya Bandi, Hari Bandi, Abir Harrasse

This paper explores optimal architectures for evaluating the outputs of large language models (LLMs) using LLMs themselves. We propose a novel framework that interprets LLMs as advocates within an ensemble of interacting agents, allowing them to defend their answers and reach conclusions through a judge and jury system. This approach offers a more dynamic and comprehensive evaluation process compared to traditional human-based assessments or automated metrics. We discuss the motivation behind this framework, its key components, and comparative advantages. We also present a probabilistic model to evaluate the error reduction achieved by iterative advocate systems. Finally, we outline experiments to validate the effectiveness of multi-advocate architectures and discuss future research directions.

☆ Contrastive Learning to Improve Retrieval for Real-world Fact Checking EMNLP 2024

Aniruddh Sriram, Fangyuan Xu, Eunsol Choi, Greg Durrett

Recent work on fact-checking addresses a realistic setting where models incorporate evidence retrieved from the web to decide the veracity of claims. A bottleneck in this pipeline is in retrieving relevant evidence: traditional methods may surface documents directly related to a claim, but fact-checking complex claims requires more inferences. For instance, a document about how a vaccine was developed is relevant to addressing claims about what it might contain, even if it does not address them directly. We present Contrastive Fact-Checking Reranker (CFR), an improved retriever for this setting. By leveraging the AVeriTeC dataset, which annotates subquestions for claims with human written answers from evidence documents, we fine-tune Contriever with a contrastive objective based on multiple training signals, including distillation from GPT-4, evaluating subquestion answers, and gold labels in the dataset. We evaluate our model on both retrieval and end-to-end veracity judgments about claims. On the AVeriTeC dataset, we find a 6\% improvement in veracity classification accuracy. We also show our gains can be transferred to FEVER, ClaimDecomp, HotpotQA, and a synthetic dataset requiring retrievers to make inferences.

comment: EMNLP 2024 FEVER Workshop

♻ ☆ Frame-Voyager: Learning to Query Frames for Video Large Language Models

Sicheng Yu, Chengkai Jin, Huanyu Wang, Zhenghao Chen, Sheng Jin, Zhongrong Zuo, Xiaolei Xu, Zhenbang Sun, Bingni Zhang, Jiawei Wu, Hao Zhang, Qianru Sun

Video Large Language Models (Video-LLMs) have made remarkable progress in video understanding tasks. However, they are constrained by the maximum length of input tokens, making it impractical to input entire videos. Existing frame selection approaches, such as uniform frame sampling and text-frame retrieval, fail to account for the information density variations in the videos or the complex instructions in the tasks, leading to sub-optimal performance. In this paper, we propose Frame-Voyager that learns to query informative frame combinations, based on the given textual queries in the task. To train Frame-Voyager, we introduce a new data collection and labeling pipeline, by ranking frame combinations using a pre-trained Video-LLM. Given a video of M frames, we traverse its T-frame combinations, feed them into a Video-LLM, and rank them based on Video-LLM's prediction losses. Using this ranking as supervision, we train Frame-Voyager to query the frame combinations with lower losses. In experiments, we evaluate Frame-Voyager on four Video Question Answering benchmarks by plugging it into two different Video-LLMs. The experimental results demonstrate that Frame-Voyager achieves impressive results in all settings, highlighting its potential as a plug-and-play solution for Video-LLMs.

comment: 19 pages, 10 figures

♻ ☆ Multi-LogiEval: Towards Evaluating Multi-Step Logical Reasoning Ability of Large Language Models EMNLP 2024

Nisarg Patel, Mohith Kulkarni, Mihir Parmar, Aashna Budhiraja, Mutsumi Nakamura, Neeraj Varshney, Chitta Baral

As Large Language Models (LLMs) continue to exhibit remarkable performance in natural language understanding tasks, there is a crucial need to measure their ability for human-like multi-step logical reasoning. Existing logical reasoning evaluation benchmarks often focus primarily on simplistic single-step or multi-step reasoning with a limited set of inference rules. Furthermore, the lack of datasets for evaluating non-monotonic reasoning represents a crucial gap since it aligns more closely with human-like reasoning. To address these limitations, we propose Multi-LogiEval, a comprehensive evaluation dataset encompassing multi-step logical reasoning with various inference rules and depths. Multi-LogiEval covers three logic types--propositional, first-order, and non-monotonic--consisting of more than 30 inference rules and more than 60 of their combinations with various depths. Leveraging this dataset, we conduct evaluations on a range of LLMs including GPT-4, ChatGPT, Gemini-Pro, Yi, Orca, and Mistral, employing a zero-shot chain-of-thought. Experimental results show that there is a significant drop in the performance of LLMs as the reasoning steps/depth increases (average accuracy of ~68% at depth-1 to ~43% at depth-5). We further conduct a thorough investigation of reasoning chains generated by LLMs which reveals several important findings. We believe that Multi-LogiEval facilitates future research for evaluating and enhancing the logical reasoning ability of LLMs. Data is available at https://github.com/Mihir3009/Multi-LogiEval.

comment: Accepted at EMNLP 2024 Main

♻ ☆ mDPO: Conditional Preference Optimization for Multimodal Large Language Models EMNLP 2024

Fei Wang, Wenxuan Zhou, James Y. Huang, Nan Xu, Sheng Zhang, Hoifung Poon, Muhao Chen

Direct preference optimization (DPO) has shown to be an effective method for large language model (LLM) alignment. Recent works have attempted to apply DPO to multimodal scenarios but have found it challenging to achieve consistent improvement. Through a comparative experiment, we identify the unconditional preference problem in multimodal preference optimization, where the model overlooks the image condition. To address this problem, we propose mDPO, a multimodal DPO objective that prevents the over-prioritization of language-only preferences by also optimizing image preference. Moreover, we introduce a reward anchor that forces the reward to be positive for chosen responses, thereby avoiding the decrease in their likelihood -- an intrinsic problem of relative preference optimization. Experiments on two multimodal LLMs of different sizes and three widely used benchmarks demonstrate that mDPO effectively addresses the unconditional preference problem in multimodal preference optimization and significantly improves model performance, particularly in reducing hallucination.

comment: Accepted to EMNLP 2024 Main Conference. Project website: https://feiwang96.github.io/mDPO

♻ ☆ SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)

Matthieu Meeus, Igor Shilov, Shubham Jain, Manuel Faysse, Marek Rei, Yves-Alexandre de Montjoye

Whether LLMs memorize their training data and what this means, from privacy leakage to detecting copyright violations -- has become a rapidly growing area of research over the last two years. In recent months, more than 10 new methods have been proposed to perform Membership Inference Attacks (MIAs) against LLMs. Contrary to traditional MIAs which rely on fixed -- but randomized -- records or models, these methods are mostly evaluated on datasets collected post-hoc. Sets of members and non-members, used to evaluate the MIA, are constructed using informed guesses after the release of a model. This lack of randomization raises concerns of a distribution shift between members and non-members. In the first part, we review the literature on MIAs against LLMs. While most work focuses on sequence-level MIAs evaluated in post-hoc setups, we show that a range of target models, motivations and units of interest have been considered in the literature. We then quantify distribution shifts present in the 6 datasets used in the literature, ranging from books to papers, using a bag of word classifier. Our analysis reveals that all of them suffer from severe distribution shifts. This challenges the validity of using such setups to measure LLM memorization and may undermine the benchmarking of recently proposed methods. Yet, all hope might not be lost. In the second part, we introduce important considerations to properly evaluate MIAs against LLMs and discuss potential ways forward: randomized test splits, injections of randomized (unique) sequences, randomized finetuning, and post-hoc control methods. While each option comes with its advantages and limitations, we believe they collectively provide solid grounds to guide the development of MIA methods and study LLM memorization. We conclude by proposing comprehensive, easy-to-use benchmarks for sequence- and document-level MIAs against LLMs.

♻ ☆ MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark (Published at NeurIPS 2024 Track Datasets and Benchmarks) NeurIPS 2024

Yubo Wang, Xueguang Ma, Ge Zhang, Yuansheng Ni, Abhranil Chandra, Shiguang Guo, Weiming Ren, Aaran Arulraj, Xuan He, Ziyan Jiang, Tianle Li, Max Ku, Kai Wang, Alex Zhuang, Rongqi Fan, Xiang Yue, Wenhu Chen

In the age of large-scale language models, benchmarks like the Massive Multitask Language Understanding (MMLU) have been pivotal in pushing the boundaries of what AI can achieve in language comprehension and reasoning across diverse domains. However, as models continue to improve, their performance on these benchmarks has begun to plateau, making it increasingly difficult to discern differences in model capabilities. This paper introduces MMLU-Pro, an enhanced dataset designed to extend the mostly knowledge-driven MMLU benchmark by integrating more challenging, reasoning-focused questions and expanding the choice set from four to ten options. Additionally, MMLU-Pro eliminates the trivial and noisy questions in MMLU. Our experimental results show that MMLU-Pro not only raises the challenge, causing a significant drop in accuracy by 16% to 33% compared to MMLU but also demonstrates greater stability under varying prompts. With 24 different prompt styles tested, the sensitivity of model scores to prompt variations decreased from 4-5% in MMLU to just 2% in MMLU-Pro. Additionally, we found that models utilizing Chain of Thought (CoT) reasoning achieved better performance on MMLU-Pro compared to direct answering, which is in stark contrast to the findings on the original MMLU, indicating that MMLU-Pro includes more complex reasoning questions. Our assessments confirm that MMLU-Pro is a more discriminative benchmark to better track progress in the field.

comment: This version has been accepted and published at NeurIPS 2024 Track Datasets and Benchmarks (Spotlight)

♻ ☆ BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions

Terry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex Gu, Zhoujun Cheng, Jiawei Liu, Qian Liu, Zijian Wang, David Lo, Binyuan Hui, Niklas Muennighoff, Daniel Fried, Xiaoning Du, Harm de Vries, Leandro Von Werra

Task automation has been greatly empowered by the recent advances in Large Language Models (LLMs) via Python code, where the tasks ranging from software engineering development to general-purpose reasoning. While current benchmarks have shown that LLMs can solve tasks using programs like human developers, the majority of their evaluations are limited to short and self-contained algorithmic tasks or standalone function calls. Solving challenging and practical requires the capability of utilizing diverse function calls as tools to efficiently implement functionalities like data analysis and web development. In addition, using multiple tools to solve a task needs compositional reasoning by accurately understanding complex instructions. Fulfilling both of these characteristics can pose a great challenge for LLMs.To assess how well LLMs can solve challenging and practical tasks via programs, we introduce BigCodeBench, a benchmark that challenges LLMs to invoke multiple function calls as tools from 139 libraries and 7 domains for 1,140 fine-grained tasks. To evaluate LLMs rigorously, each task encompasses 5.6 test cases with an average branch coverage of 99%. In addition, we propose a natural-language-oriented variant of BigCodeBench, BigCodeBench-Instruct, that automatically transforms the original docstrings into short instructions only with essential information. Our extensive evaluation of 60 LLMs shows that LLMs are not yet capable of following complex instructions to use function calls precisely, with scores up to 60%, significantly lower than the human performance of 97%. The results underscore the need for further advancements in this area.

comment: 44 pages, 14 figures, 7 tables, built with love by the BigCode community :)

♻ ☆ Augmenting Black-box LLMs with Medical Textbooks for Biomedical Question Answering (Published in Findings of EMNLP 2024) EMNLP

Yubo Wang, Xueguang Ma, Wenhu Chen

Large-scale language models (LLMs) like ChatGPT have demonstrated impressive abilities in generating responses based on human instructions. However, their use in the medical field can be challenging due to their lack of specific, in-depth knowledge. In this study, we present a system called LLMs Augmented with Medical Textbooks (LLM-AMT) designed to enhance the proficiency of LLMs in specialized domains. LLM-AMT integrates authoritative medical textbooks into the LLMs' framework using plug-and-play modules. These modules include a Query Augmenter, a Hybrid Textbook Retriever, and a Knowledge Self-Refiner. Together, they incorporate authoritative medical knowledge. Additionally, an LLM Reader aids in contextual understanding. Our experimental results on three medical QA tasks demonstrate that LLMAMT significantly improves response quality, with accuracy gains ranging from 11.6% to 16.6%. Notably, with GPT-4-Turbo as the base model, LLM-AMT outperforms the specialized Med-PaLM 2 model pre-trained on a massive amount of medical corpus by 2-3%. We found that despite being 100x smaller in size, medical textbooks as a retrieval corpus is proven to be a more effective knowledge database than Wikipedia in the medical domain, boosting performance by 7.8%-13.7%.

comment: This version has been accepted and published at EMNLP Findings 2024

♻ ☆ Annotation alignment: Comparing LLM and human annotations of conversational safety EMNLP 2024

Rajiv Movva, Pang Wei Koh, Emma Pierson

Do LLMs align with human perceptions of safety? We study this question via annotation alignment, the extent to which LLMs and humans agree when annotating the safety of user-chatbot conversations. We leverage the recent DICES dataset (Aroyo et al., 2023), in which 350 conversations are each rated for safety by 112 annotators spanning 10 race-gender groups. GPT-4 achieves a Pearson correlation of $r = 0.59$ with the average annotator rating, \textit{higher} than the median annotator's correlation with the average ($r=0.51$). We show that larger datasets are needed to resolve whether LLMs exhibit disparities in how well they correlate with different demographic groups. Also, there is substantial idiosyncratic variation in correlation within groups, suggesting that race & gender do not fully capture differences in alignment. Finally, we find that GPT-4 cannot predict when one demographic group finds a conversation more unsafe than another.

comment: EMNLP 2024 (Main). Main text contains 6 pages, 2 figures

♻ ☆ Contextual Document Embeddings

John X. Morris, Alexander M. Rush

Dense document embeddings are central to neural retrieval. The dominant paradigm is to train and construct embeddings by running encoders directly on individual documents. In this work, we argue that these embeddings, while effective, are implicitly out-of-context for targeted use cases of retrieval, and that a contextualized document embedding should take into account both the document and neighboring documents in context - analogous to contextualized word embeddings. We propose two complementary methods for contextualized document embeddings: first, an alternative contrastive learning objective that explicitly incorporates the document neighbors into the intra-batch contextual loss; second, a new contextual architecture that explicitly encodes neighbor document information into the encoded representation. Results show that both methods achieve better performance than biencoders in several settings, with differences especially pronounced out-of-domain. We achieve state-of-the-art results on the MTEB benchmark with no hard negative mining, score distillation, dataset-specific instructions, intra-GPU example-sharing, or extremely large batch sizes. Our method can be applied to improve performance on any contrastive learning dataset and any biencoder.

♻ ☆ Creative Beam Search: LLM-as-a-Judge For Improving Response Generation

Giorgio Franceschelli, Mirco Musolesi

Large language models are revolutionizing several areas, including artificial creativity. However, the process of generation in machines profoundly diverges from that observed in humans. In particular, machine generation is characterized by a lack of intentionality and an underlying creative process. We propose a method called Creative Beam Search that uses Diverse Beam Search and LLM-as-a-Judge to perform response generation and response validation. The results of a qualitative experiment show how our approach can provide better output than standard sampling techniques. We also show that the response validation step is a necessary complement to the response generation step.

comment: Presented as a short paper at the 15th International Conference on Computational Creativity (ICCC'24)

♻ ☆ MetaMetrics: Calibrating Metrics For Generation Tasks Using Human Preferences

Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, Derry Tanti Wijaya

Understanding the quality of a performance evaluation metric is crucial for ensuring that model outputs align with human preferences. However, it remains unclear how well each metric captures the diverse aspects of these preferences, as metrics often excel in one particular area but not across all dimensions. To address this, it is essential to systematically calibrate metrics to specific aspects of human preference, catering to the unique characteristics of each aspect. We introduce MetaMetrics, a calibrated meta-metric designed to evaluate generation tasks across different modalities in a supervised manner. MetaMetrics optimizes the combination of existing metrics to enhance their alignment with human preferences. Our metric demonstrates flexibility and effectiveness in both language and vision downstream tasks, showing significant benefits across various multilingual and multi-domain scenarios. MetaMetrics aligns closely with human preferences and is highly extendable and easily integrable into any application. This makes MetaMetrics a powerful tool for improving the evaluation of generation tasks, ensuring that metrics are more representative of human judgment across diverse contexts.

comment: Preprint

♻ ☆ A Usage-centric Take on Intent Understanding in E-Commerce EMNLP 2024

Wendi Zhou, Tianyi Li, Pavlos Vougiouklis, Mark Steedman, Jeff Z. Pan

Identifying and understanding user intents is a pivotal task for E-Commerce. Despite its essential role in product recommendation and business user profiling analysis, intent understanding has not been consistently defined or accurately benchmarked. In this paper, we focus on predicative user intents as "how a customer uses a product", and pose intent understanding as a natural language reasoning task, independent of product ontologies. We identify two weaknesses of FolkScope, the SOTA E-Commerce Intent Knowledge Graph: category-rigidity and property-ambiguity. They limit its ability to strongly align user intents with products having the most desirable property, and to recommend useful products across diverse categories. Following these observations, we introduce a Product Recovery Benchmark featuring a novel evaluation framework and an example dataset. We further validate the above FolkScope weaknesses on this benchmark. Our code and dataset are available at https://github.com/stayones/Usgae-Centric-Intent-Understanding.

comment: Acepted by EMNLP 2024 main

♻ ☆ Efficient Model-Agnostic Multi-Group Equivariant Networks

Razan Baltaji, Sourya Basu, Lav R. Varshney

Constructing model-agnostic group equivariant networks, such as equitune (Basu et al., 2023b) and its generalizations (Kim et al., 2023), can be computationally expensive for large product groups. We address this problem by providing efficient model-agnostic equivariant designs for two related problems: one where the network has multiple inputs each with potentially different groups acting on them, and another where there is a single input but the group acting on it is a large product group. For the first design, we initially consider a linear model and characterize the entire equivariant space that satisfies this constraint. This characterization gives rise to a novel fusion layer between different channels that satisfies an invariance-symmetry (IS) constraint, which we call an IS layer. We then extend this design beyond linear models, similar to equitune, consisting of equivariant and IS layers. We also show that the IS layer is a universal approximator of invariant-symmetric functions. Inspired by the first design, we use the notion of the IS property to design a second efficient model-agnostic equivariant design for large product groups acting on a single input. For the first design, we provide experiments on multi-image classification where each view is transformed independently with transformations such as rotations. We find equivariant models are robust to such transformations and perform competitively otherwise. For the second design, we consider three applications: language compositionality on the SCAN dataset to product groups; fairness in natural language generation from GPT-2 to address intersectionality; and robust zero-shot image classification with CLIP. Overall, our methods are simple and general, competitive with equitune and its variants, while also being computationally more efficient.

♻ ☆ When "A Helpful Assistant" Is Not Really Helpful: Personas in System Prompts Do Not Improve Performances of Large Language Models

Mingqian Zheng, Jiaxin Pei, Lajanugen Logeswaran, Moontae Lee, David Jurgens

Prompting serves as the major way humans interact with Large Language Models (LLM). Commercial AI systems commonly define the role of the LLM in system prompts. For example, ChatGPT uses "You are a helpful assistant" as part of its default system prompt. Despite current practices of adding personas to system prompts, it remains unclear how different personas affect a model's performance on objective tasks. In this study, we present a systematic evaluation of personas in system prompts. We curate a list of 162 roles covering 6 types of interpersonal relationships and 8 domains of expertise. Through extensive analysis of 4 popular families of LLMs and 2,410 factual questions, we demonstrate that adding personas in system prompts does not improve model performance across a range of questions compared to the control setting where no persona is added. Nevertheless, further analysis suggests that the gender, type, and domain of the persona can all influence the resulting prediction accuracies. We further experimented with a list of persona search strategies and found that, while aggregating results from the best persona for each question significantly improves prediction accuracy, automatically identifying the best persona is challenging, with predictions often performing no better than random selection. Overall, our findings suggest that while adding a persona may lead to performance gains in certain settings, the effect of each persona can be largely random. Code and data are available at https://github.com/Jiaxin-Pei/Prompting-with-Social-Roles.

♻ ☆ Better Instruction-Following Through Minimum Bayes Risk

Ian Wu, Patrick Fernandes, Amanda Bertsch, Seungone Kim, Sina Pakazad, Graham Neubig

General-purpose LLM judges capable of human-level evaluation provide not only a scalable and accurate way of evaluating instruction-following LLMs but also new avenues for supervising and improving their performance. One promising way of leveraging LLM judges for supervision is through Minimum Bayes Risk (MBR) decoding, which uses a reference-based evaluator to select a high-quality output from amongst a set of candidate outputs. In the first part of this work, we explore using MBR decoding as a method for improving the test-time performance of instruction-following LLMs. We find that MBR decoding with reference-based LLM judges substantially improves over greedy decoding, best-of-N decoding with reference-free judges and MBR decoding with lexical and embedding-based metrics on AlpacaEval and MT-Bench. These gains are consistent across LLMs with up to 70B parameters, demonstrating that smaller LLM judges can be used to supervise much larger LLMs. Then, seeking to retain the improvements from MBR decoding while mitigating additional test-time costs, we explore iterative self-training on MBR-decoded outputs. We find that self-training using Direct Preference Optimisation leads to significant performance gains, such that the self-trained models with greedy decoding generally match and sometimes exceed the performance of their base models with MBR decoding.

♻ ☆ Representation noising effectively prevents harmful fine-tuning on LLMs

Domenic Rosati, Jan Wehner, Kai Williams, Łukasz Bartoszcze, David Atanasov, Robie Gonzales, Subhabrata Majumdar, Carsten Maple, Hassan Sajjad, Frank Rudzicz

Releasing open-source large language models (LLMs) presents a dual-use risk since bad actors can easily fine-tune these models for harmful purposes. Even without the open release of weights, weight stealing and fine-tuning APIs make closed models vulnerable to harmful fine-tuning attacks (HFAs). While safety measures like preventing jailbreaks and improving safety guardrails are important, such measures can easily be reversed through fine-tuning. In this work, we propose Representation Noising (RepNoise), a defence mechanism that is effective even when attackers have access to the weights. RepNoise works by removing information about harmful representations such that it is difficult to recover them during fine-tuning. Importantly, our defence is also able to generalize across different subsets of harm that have not been seen during the defence process as long as they are drawn from the same distribution of the attack set. Our method does not degrade the general capability of LLMs and retains the ability to train the model on harmless tasks. We provide empirical evidence that the effectiveness of our defence lies in its "depth": the degree to which information about harmful representations is removed across all layers of the LLM.

comment: Published in NeurIPs 2024

♻ ☆ Social Bias Probing: Fairness Benchmarking for Language Models

Marta Marchiori Manerba, Karolina Stańczak, Riccardo Guidotti, Isabelle Augenstein

While the impact of social biases in language models has been recognized, prior methods for bias evaluation have been limited to binary association tests on small datasets, limiting our understanding of bias complexities. This paper proposes a novel framework for probing language models for social biases by assessing disparate treatment, which involves treating individuals differently according to their affiliation with a sensitive demographic group. We curate SoFa, a large-scale benchmark designed to address the limitations of existing fairness collections. SoFa expands the analysis beyond the binary comparison of stereotypical versus anti-stereotypical identities to include a diverse range of identities and stereotypes. Comparing our methodology with existing benchmarks, we reveal that biases within language models are more nuanced than acknowledged, indicating a broader scope of encoded biases than previously recognized. Benchmarking LMs on SoFa, we expose how identities expressing different religions lead to the most pronounced disparate treatments across all models. Finally, our findings indicate that real-life adversities faced by various groups such as women and people with disabilities are mirrored in the behavior of these models.

♻ ☆ Fine-Tuning and Prompt Optimization: Two Great Steps that Work Better Together EMNLP 2024

Dilara Soylu, Christopher Potts, Omar Khattab

Natural Language Processing (NLP) systems are increasingly taking the form of sophisticated modular pipelines, e.g., Retrieval Augmented Generation (RAG), where each module may involve a distinct Language Model (LM) and an associated prompt template. These compound systems often lack intermediate labels or gradient flow to optimize each module, making their end-to-end optimization challenging. Here we seek strategies to optimize both the module-level LM weights and the associated prompt templates of such systems to maximize a downstream task metric. We propose for the first time combining the weight and prompt optimization strategies to optimize a modular LM pipeline by alternating between the two to get the same LM to teach itself. In experiments with multi-hop QA, mathematical reasoning, and feature-based classification using mistral-7b, llama-2-7b, and llama-3-8b, these BetterTogether strategies optimizing the weights and prompts of a pipeline together outperform directly optimizing weights alone and prompts alone by up to 60% and 6%, respectively, on average across LMs and tasks. BetterTogether optimizer is released in DSPy at http://dspy.ai

comment: EMNLP 2024

♻ ☆ FAC$^2$E: Better Understanding Large Language Model Capabilities by Dissociating Language and Cognition EMNLP 2024

Xiaoqiang Wang, Lingfei Wu, Tengfei Ma, Bang Liu

Large language models (LLMs) are primarily evaluated by overall performance on various text understanding and generation tasks. However, such a paradigm fails to comprehensively differentiate the fine-grained language and cognitive skills, rendering the lack of sufficient interpretation to LLMs' capabilities. In this paper, we present FAC$^2$E, a framework for Fine-grAined and Cognition-grounded LLMs' Capability Evaluation. Specifically, we formulate LLMs' evaluation in a multi-dimensional and explainable manner by dissociating the language-related capabilities and the cognition-related ones. Besides, through extracting the intermediate reasoning from LLMs, we further break down the process of applying a specific capability into three sub-steps: recalling relevant knowledge, utilizing knowledge, and solving problems. Finally, FAC$^2$E evaluates each sub-step of each fine-grained capability, providing a two-faceted diagnosis for LLMs. Utilizing FAC$^2$E, we identify a common shortfall in knowledge utilization among models and propose a straightforward, knowledge-enhanced method to mitigate this issue. Our results not only showcase promising performance enhancements but also highlight a direction for future LLM advancements.

comment: Accepted at EMNLP 2024 main conference

♻ ☆ Non-Invasive Suicide Risk Prediction Through Speech Analysis

Shahin Amiriparian, Maurice Gerczuk, Justina Lutz, Wolfgang Strube, Irina Papazova, Alkomiet Hasan, Alexander Kathan, Björn W. Schuller

The delayed access to specialized psychiatric assessments and care for patients at risk of suicidal tendencies in emergency departments creates a notable gap in timely intervention, hindering the provision of adequate mental health support during critical situations. To address this, we present a non-invasive, speech-based approach for automatic suicide risk assessment. For our study, we collected a novel speech recording dataset from $20$ patients. We extract three sets of features, including wav2vec, interpretable speech and acoustic features, and deep learning-based spectral representations. We proceed by conducting a binary classification to assess suicide risk in a leave-one-subject-out fashion. Our most effective speech model achieves a balanced accuracy of $66.2\,\%$. Moreover, we show that integrating our speech model with a series of patients' metadata, such as the history of suicide attempts or access to firearms, improves the overall result. The metadata integration yields a balanced accuracy of $94.4\,\%$, marking an absolute improvement of $28.2\,\%$, demonstrating the efficacy of our proposed approaches for automatic suicide risk assessment in emergency medicine.

♻ ☆ Native Design Bias: Studying the Impact of English Nativeness on Language Model Performance

Manon Reusens, Philipp Borchert, Jochen De Weerdt, Bart Baesens

Large Language Models (LLMs) excel at providing information acquired during pretraining on large-scale corpora and following instructions through user prompts. This study investigates whether the quality of LLM responses varies depending on the demographic profile of users. Considering English as the global lingua franca, along with the diversity of its dialects among speakers of different native languages, we explore whether non-native English speakers receive lower-quality or even factually incorrect responses from LLMs more frequently. Our results show that performance discrepancies occur when LLMs are prompted by native versus non-native English speakers and persist when comparing native speakers from Western countries with others. Additionally, we find a strong anchoring effect when the model recognizes or is made aware of the user's nativeness, which further degrades the response quality when interacting with non-native speakers. Our analysis is based on a newly collected dataset with over 12,000 unique annotations from 124 annotators, including information on their native language and English proficiency.

♻ ☆ UPCS: Unbiased Persona Construction for Dialogue Generation

Kuiyun Chen, Yanbin Wei

Narrative systems, such as dialogue and storytelling systems, often utilize persona profiles to enhance personalized interactions. Existing persona profiles frequently exhibit biases, posing risks to system integrity and fairness. To address this, we introduce the UPCS framework, which categorizes character descriptions into eight dimensions, including bias mitigation strategies. Experimental results demonstrate UPCS's superiority in accuracy, diversity, bias elimination, and user satisfaction, marking a significant advancement in persona construction for reliable narrative systems.

♻ ☆ Diversity Over Size: On the Effect of Sample and Topic Sizes for Topic-Dependent Argument Mining Datasets

Benjamin Schiller, Johannes Daxenberger, Andreas Waldis, Iryna Gurevych

The task of Argument Mining, that is extracting and classifying argument components for a specific topic from large document sources, is an inherently difficult task for machine learning models and humans alike, as large Argument Mining datasets are rare and recognition of argument components requires expert knowledge. The task becomes even more difficult if it also involves stance detection of retrieved arguments. In this work, we investigate the effect of Argument Mining dataset composition in few- and zero-shot settings. Our findings show that, while fine-tuning is mandatory to achieve acceptable model performance, using carefully composed training samples and reducing the training sample size by up to almost 90% can still yield 95% of the maximum performance. This gain is consistent across three Argument Mining tasks on three different datasets. We also publish a new dataset for future benchmarking.

♻ ☆ KV-Compress: Paged KV-Cache Compression with Variable Compression Rates per Attention Head

Isaac Rehg

Context lengths of Large Language Models (LLMs) have exploded in recent years, with 128k-token context becoming a standard and million-token context becoming a reality. Efficiently supporting long-context inference remains challenging as the memory that must be allocated in key-value (KV) cache for a generation scales with its context length, limiting the number of long-context requests that can be served concurrently under a given memory budget. KV cache compression can mitigate this issue by removing under-utilized KVs from each attention head's cache and reducing its memory footprint. Higher theoretical compression rates can be achieved when the number of removed KVs varies across attention heads, but application of such a strategy within existing inference frameworks adds fragmentation and cannot realize the theoretical compression rates in physical memory. We introduce KV-Compress, a novel compression method that evicts contiguous KV blocks within a PagedAttention framework, reducing the memory footprint of the KV cache proportionally to this theoretical compression rate. Our method achieves state-of-the-art performance on LongBench for both Mistral-7B-Instruct-v0.2 and Llama-3.1-8B-Instruct while lowering the total number of compressed KVs by 4x compared with prior methods. Evaluations on Llama-3.1-8B-Instruct and Llama-3.1-70B-Instruct-FP8 achieve compression rates up to 8x with negligible impact on performance, and up to 64x while retaining over 90% of full-cache performance for all but three of the suite's subsets. We benchmark an integration of our method with vLLM that increases total throughput by up to 5.18x by enabling larger decoding batches.

♻ ☆ Decoding Intelligence: A Framework for Certifying Knowledge Comprehension in LLMs

Isha Chaudhary, Vedaant V. Jain, Gagandeep Singh

Knowledge comprehension capability is an important aspect of human intelligence. As Large Language Models (LLMs) are being envisioned as superhuman agents, it is crucial for them to be proficient at knowledge comprehension. However, existing benchmarking studies do not provide consistent, generalizable, and formal guarantees on the knowledge comprehension capabilities of LLMs. In this work, we propose the first framework to certify knowledge comprehension in LLMs with formal probabilistic guarantees. Our certificates are quantitative -- they consist of high-confidence, tight bounds on the probability that a target LLM gives the correct answer on any knowledge comprehension prompt sampled from a distribution. We design and certify novel specifications that precisely represent distributions of knowledge comprehension prompts leveraging knowledge graphs. We certify SOTA LLMs for specifications over the Wikidata5m knowledge graph. We find that the knowledge comprehension capability improves significantly with scaling the size of the models.

♻ ☆ First Heuristic Then Rational: Dynamic Use of Heuristics in Language Model Reasoning EMNLP 2024

Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Keisuke Sakaguchi, Kentaro Inui

Multi-step reasoning instruction, such as chain-of-thought prompting, is widely adopted to explore better language models (LMs) performance. We report on the systematic strategy that LMs employ in such a multi-step reasoning process. Our controlled experiments reveal that LMs rely more heavily on heuristics, such as lexical overlap, in the earlier stages of reasoning, where more reasoning steps remain to reach a goal. Conversely, their reliance on heuristics decreases as LMs progress closer to the final answer through multiple reasoning steps. This suggests that LMs can backtrack only a limited number of future steps and dynamically combine heuristic strategies with rationale ones in tasks involving multi-step reasoning.

comment: This paper is accepted at EMNLP 2024

♻ ☆ LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations

Hadas Orgad, Michael Toker, Zorik Gekhman, Roi Reichart, Idan Szpektor, Hadas Kotek, Yonatan Belinkov

Large language models (LLMs) often produce errors, including factual inaccuracies, biases, and reasoning failures, collectively referred to as "hallucinations". Recent studies have demonstrated that LLMs' internal states encode information regarding the truthfulness of their outputs, and that this information can be utilized to detect errors. In this work, we show that the internal representations of LLMs encode much more information about truthfulness than previously recognized. We first discover that the truthfulness information is concentrated in specific tokens, and leveraging this property significantly enhances error detection performance. Yet, we show that such error detectors fail to generalize across datasets, implying that -- contrary to prior claims -- truthfulness encoding is not universal but rather multifaceted. Next, we show that internal representations can also be used for predicting the types of errors the model is likely to make, facilitating the development of tailored mitigation strategies. Lastly, we reveal a discrepancy between LLMs' internal encoding and external behavior: they may encode the correct answer, yet consistently generate an incorrect one. Taken together, these insights deepen our understanding of LLM errors from the model's internal perspective, which can guide future research on enhancing error analysis and mitigation.

♻ ☆ StructLM: Towards Building Generalist Models for Structured Knowledge Grounding

Alex Zhuang, Ge Zhang, Tianyu Zheng, Xinrun Du, Junjie Wang, Weiming Ren, Stephen W. Huang, Jie Fu, Xiang Yue, Wenhu Chen

Structured data sources, such as tables, graphs, and databases, are ubiquitous knowledge sources. Despite the demonstrated capabilities of large language models (LLMs) on plain text, their proficiency in interpreting and utilizing structured data remains limited. Our investigation reveals a notable deficiency in LLMs' ability to process structured data, e.g., ChatGPT lags behind state-of-the-art (SoTA) model by an average of 35%. To augment the Structured Knowledge Grounding (SKG) capabilities in LLMs, we have developed a comprehensive instruction tuning dataset comprising 1.1 million examples. Utilizing this dataset, we train a series of models, referred to as StructLM, based on the Mistral and the CodeLlama model family, ranging from 7B to 34B parameters. Our StructLM series surpasses task-specific models on 16 out of 18 evaluated datasets and establishes new SoTA performance on 8 SKG tasks. Furthermore, StructLM demonstrates strong generalization across 6 novel held-out SKG tasks, outperforming TableLlama by an average of 35\% and Flan-UL2 20B by an average of 10\%. Contrary to expectations, we observe that scaling model size offers marginal benefits, with StructLM-34B showing only slight improvements over StructLM-7B. This suggests that structured knowledge grounding is still a challenging task and requires more innovative design to push to a new level.

comment: Technical Report

♻ ☆ WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models NeurIPS 2024

Peng Wang, Zexi Li, Ningyu Zhang, Ziwen Xu, Yunzhi Yao, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen

Large language models (LLMs) need knowledge updates to meet the ever-growing world facts and correct the hallucinated responses, facilitating the methods of lifelong model editing. Where the updated knowledge resides in memories is a fundamental question for model editing. In this paper, we find that editing either long-term memory (direct model parameters) or working memory (non-parametric knowledge of neural network activations/representations by retrieval) will result in an impossible triangle -- reliability, generalization, and locality can not be realized together in the lifelong editing settings. For long-term memory, directly editing the parameters will cause conflicts with irrelevant pretrained knowledge or previous edits (poor reliability and locality). For working memory, retrieval-based activations can hardly make the model understand the edits and generalize (poor generalization). Therefore, we propose WISE to bridge the gap between memories. In WISE, we design a dual parametric memory scheme, which consists of the main memory for the pretrained knowledge and a side memory for the edited knowledge. We only edit the knowledge in the side memory and train a router to decide which memory to go through when given a query. For continual editing, we devise a knowledge-sharding mechanism where different sets of edits reside in distinct subspaces of parameters, and are subsequently merged into a shared memory without conflicts. Extensive experiments show that WISE can outperform previous model editing methods and overcome the impossible triangle under lifelong model editing of question answering, hallucination, and out-of-distribution settings across trending LLM architectures, e.g., GPT, LLaMA, and Mistral. Code is available at https://github.com/zjunlp/EasyEdit.

comment: NeurIPS 2024

♻ ☆ Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts

Junmo Kang, Leonid Karlinsky, Hongyin Luo, Zhen Wang, Jacob Hansen, James Glass, David Cox, Rameswar Panda, Rogerio Feris, Alan Ritter

We present Self-MoE, an approach that transforms a monolithic LLM into a compositional, modular system of self-specialized experts, named MiXSE (MiXture of Self-specialized Experts). Our approach leverages self-specialization, which constructs expert modules using self-generated synthetic data, each equipping a shared base LLM with distinct domain-specific capabilities, activated via self-optimized routing. This allows for dynamic and capability-specific handling of various target tasks, enhancing overall capabilities, without extensive human-labeled data and added parameters. Our empirical results reveal that specializing LLMs may exhibit potential trade-offs in performances on non-specialized tasks. On the other hand, our Self-MoE demonstrates substantial improvements (6.5%p on average) over the base LLM across diverse benchmarks such as knowledge, reasoning, math, and coding. It also consistently outperforms other methods, including instance merging and weight merging, while offering better flexibility and interpretability by design with semantic experts and routing. Our findings highlight the critical role of modularity, the applicability of Self-MoE to multiple base LLMs, and the potential of self-improvement in achieving efficient, scalable, and adaptable systems.

♻ ☆ A Russian Jeopardy! Data Set for Question-Answering Systems

Elena Mikhalkova

Question answering (QA) is one of the most common NLP tasks that relates to named entity recognition, fact extraction, semantic search and some other fields. In industry, it is much appreciated in chatbots and corporate information systems. It is also a challenging task that attracted the attention of a very general audience at the quiz show Jeopardy! In this article we describe a Jeopardy!-like Russian QA data set collected from the official Russian quiz database Chgk (che ge ka). The data set includes 379,284 quiz-like questions with 29,375 from the Russian analogue of Jeopardy! - "Own Game". We observe its linguistic features and the related QA-task. We conclude about perspectives of a QA competition based on the data set collected from this database.

♻ ☆ PEAR: Position-Embedding-Agnostic Attention Re-weighting Enhances Retrieval-Augmented Generation with Zero Inference Overhead

Tao Tan, Yining Qian, Ang Lv, Hongzhan Lin, Songhao Wu, Yongbo Wang, Feng Wang, Jingtong Wu, Xin Lu, Rui Yan

Large language models (LLMs) enhanced with retrieval-augmented generation (RAG) have introduced a new paradigm for web search. However, the limited context awareness of LLMs degrades their performance on RAG tasks. Existing methods to enhance context awareness are often inefficient, incurring time or memory overhead during inference, and many are tailored to specific position embeddings. In this paper, we propose Position-Embedding-Agnostic attention Re-weighting (PEAR), which enhances the context awareness of LLMs with zero inference overhead. Specifically, on a proxy task focused on context copying, we first detect heads which suppress the models' context awareness thereby diminishing RAG performance. To weaken the impact of these heads, we re-weight their outputs with learnable coefficients. The LLM (with frozen parameters) is optimized by adjusting these coefficients to minimize loss on the proxy task. As a result, the coefficients are optimized to values less than one, thereby reducing their tendency to suppress RAG performance. During inference, the optimized coefficients are fixed to re-weight these heads, regardless of the specific task at hand. Our proposed PEAR offers two major advantages over previous approaches: (1) It introduces zero additional inference overhead in terms of memory usage or inference time, while outperforming competitive baselines in accuracy and efficiency across various RAG tasks. (2) It is independent of position embedding algorithms, ensuring broader applicability.

comment: preprint

♻ ☆ WellDunn: On the Robustness and Explainability of Language Models and Large Language Models in Identifying Wellness Dimensions EMNLP 2024

Seyedali Mohammadi, Edward Raff, Jinendra Malekar, Vedant Palit, Francis Ferraro, Manas Gaur

Language Models (LMs) are being proposed for mental health applications where the heightened risk of adverse outcomes means predictive performance may not be a sufficient litmus test of a model's utility in clinical practice. A model that can be trusted for practice should have a correspondence between explanation and clinical determination, yet no prior research has examined the attention fidelity of these models and their effect on ground truth explanations. We introduce an evaluation design that focuses on the robustness and explainability of LMs in identifying Wellness Dimensions (WDs). We focus on two existing mental health and well-being datasets: (a) Multi-label Classification-based MultiWD, and (b) WellXplain for evaluating attention mechanism veracity against expert-labeled explanations. The labels are based on Halbert Dunn's theory of wellness, which gives grounding to our evaluation. We reveal four surprising results about LMs/LLMs: (1) Despite their human-like capabilities, GPT-3.5/4 lag behind RoBERTa, and MedAlpaca, a fine-tuned LLM on WellXplain fails to deliver any remarkable improvements in performance or explanations. (2) Re-examining LMs' predictions based on a confidence-oriented loss function reveals a significant performance drop. (3) Across all LMs/LLMs, the alignment between attention and explanations remains low, with LLMs scoring a dismal 0.0. (4) Most mental health-specific LMs/LLMs overlook domain-specific knowledge and undervalue explanations, causing these discrepancies. This study highlights the need for further research into their consistency and explanations in mental health and well-being.

comment: Accepted in BlackboxNLP @ EMNLP 2024

♻ ☆ When Can Transformers Count to n?

Gilad Yehudai, Haim Kaplan, Asma Ghandeharioun, Mor Geva, Amir Globerson

Large language models based on the transformer architectures can solve highly complex tasks. But are there simple tasks that such models cannot solve? Here we focus on very simple counting tasks, that involve counting how many times a token in the vocabulary have appeared in a string. We show that if the dimension of the transformer state is linear in the context length, this task can be solved. However, the solution we propose does not scale beyond this limit, and we provide theoretical arguments for why it is likely impossible for a size limited transformer to implement this task. Our empirical results demonstrate the same phase-transition in performance, as anticipated by the theoretical argument. Our results demonstrate the importance of understanding how transformers can solve simple tasks.

♻ ☆ Tokenization Is More Than Compression EMNLP 2024

Craig W. Schmidt, Varshini Reddy, Haoran Zhang, Alec Alameddine, Omri Uzan, Yuval Pinter, Chris Tanner

Tokenization is a foundational step in natural language processing (NLP) tasks, bridging raw text and language models. Existing tokenization approaches like Byte-Pair Encoding (BPE) originate from the field of data compression, and it has been suggested that the effectiveness of BPE stems from its ability to condense text into a relatively small number of tokens. We test the hypothesis that fewer tokens lead to better downstream performance by introducing PathPiece, a new tokenizer that segments a document's text into the minimum number of tokens for a given vocabulary. Through extensive experimentation we find this hypothesis not to be the case, casting doubt on the understanding of the reasons for effective tokenization. To examine which other factors play a role, we evaluate design decisions across all three phases of tokenization: pre-tokenization, vocabulary construction, and segmentation, offering new insights into the design of effective tokenizers. Specifically, we illustrate the importance of pre-tokenization and the benefits of using BPE to initialize vocabulary construction. We train 64 language models with varying tokenization, ranging in size from 350M to 2.4B parameters, all of which are made publicly available.

comment: EMNLP 2024

♻ ☆ ComplexTempQA: A Large-Scale Dataset for Complex Temporal Question Answering

Raphael Gruber, Abdelrahman Abdallah, Michael Färber, Adam Jatowt

We introduce ComplexTempQA, a large-scale dataset consisting of over 100 million question-answer pairs designed to tackle the challenges in temporal question answering. ComplexTempQA significantly surpasses existing benchmarks like HOTPOTQA, TORQUE, and TEQUILA in scale and scope. Utilizing data from Wikipedia and Wikidata, the dataset covers questions spanning over two decades and offers an unmatched breadth of topics. We introduce a unique taxonomy that categorizes questions as attributes, comparisons, and counting questions, each revolving around events, entities, and time periods. One standout feature of ComplexTempQA is the high complexity of its questions, which demand effective capabilities for answering such as across-time comparison, temporal aggregation, and multi-hop reasoning involving temporal event ordering and entity recognition. Additionally, each question is accompanied by detailed metadata, including specific time scopes, allowing for comprehensive evaluation and enhancement of the temporal reasoning abilities of large language models. ComplexTempQA serves both as a testing ground for developing sophisticated AI models and as a foundation for advancing research in question answering, information retrieval, and language understanding.

♻ ☆ Reconstruct Your Previous Conversations! Comprehensively Investigating Privacy Leakage Risks in Conversations with GPT Models EMNLP 2024

Junjie Chu, Zeyang Sha, Michael Backes, Yang Zhang

Significant advancements have recently been made in large language models represented by GPT models. Users frequently have multi-round private conversations with cloud-hosted GPT models for task optimization. Yet, this operational paradigm introduces additional attack surfaces, particularly in custom GPTs and hijacked chat sessions. In this paper, we introduce a straightforward yet potent Conversation Reconstruction Attack. This attack targets the contents of previous conversations between GPT models and benign users, i.e., the benign users' input contents during their interaction with GPT models. The adversary could induce GPT models to leak such contents by querying them with designed malicious prompts. Our comprehensive examination of privacy risks during the interactions with GPT models under this attack reveals GPT-4's considerable resilience. We present two advanced attacks targeting improved reconstruction of past conversations, demonstrating significant privacy leakage across all models under these advanced techniques. Evaluating various defense mechanisms, we find them ineffective against these attacks. Our findings highlight the ease with which privacy can be compromised in interactions with GPT models, urging the community to safeguard against potential abuses of these models' capabilities.

comment: Accepted in EMNLP 2024. 14 pages, 10 figures

♻ ☆ Visual Question Decomposition on Multimodal Large Language Models EMNLP2024

Haowei Zhang, Jianzhe Liu, Zhen Han, Shuo Chen, Bailan He, Volker Tresp, Zhiqiang Xu, Jindong Gu

Question decomposition has emerged as an effective strategy for prompting Large Language Models (LLMs) to answer complex questions. However, while existing methods primarily focus on unimodal language models, the question decomposition capability of Multimodal Large Language Models (MLLMs) has yet to be explored. To this end, this paper explores visual question decomposition on MLLMs. Specifically, we introduce a systematic evaluation framework including a dataset and several evaluation criteria to assess the quality of the decomposed sub-questions, revealing that existing MLLMs struggle to produce high-quality sub-questions. To address this limitation, we propose a specific finetuning dataset, DecoVQA+, for enhancing the model's question decomposition capability. Aiming at enabling models to perform appropriate selective decomposition, we propose an efficient finetuning pipeline. The finetuning pipeline consists of our proposed dataset and a training objective for selective decomposition. Finetuned MLLMs demonstrate significant improvements in the quality of sub-questions and the policy of selective question decomposition. Additionally, the models also achieve higher accuracy with selective decomposition on VQA benchmark datasets.

comment: Accepted to EMNLP2024 Findings

♻ ☆ DYNAMICQA: Tracing Internal Knowledge Conflicts in Language Models EMNLP 2024

Sara Vera Marjanović, Haeun Yu, Pepa Atanasova, Maria Maistro, Christina Lioma, Isabelle Augenstein

Knowledge-intensive language understanding tasks require Language Models (LMs) to integrate relevant context, mitigating their inherent weaknesses, such as incomplete or outdated knowledge. However, conflicting knowledge can be present in the LM's parameters, termed intra-memory conflict, which can affect a model's propensity to accept contextual knowledge. To study the effect of intra-memory conflict on an LM's ability to accept relevant context, we utilize two knowledge conflict measures and a novel dataset containing inherently conflicting data, DynamicQA. This dataset includes facts with a temporal dynamic nature where facts can change over time and disputable dynamic facts, which can change depending on the viewpoint. DynamicQA is the first to include real-world knowledge conflicts and provide context to study the link between the different types of knowledge conflicts. We also evaluate several measures on their ability to reflect the presence of intra-memory conflict: semantic entropy and a novel coherent persuasion score. With our extensive experiments, we verify that LMs exhibit a greater degree of intra-memory conflict with dynamic facts compared to facts that have a single truth value. Furthermore, we reveal that facts with intra-memory conflict are harder to update with context, suggesting that retrieval-augmented generation will struggle with the most commonly adapted facts.

comment: 15 pages, 6 figures, Accepted to Findings of EMNLP 2024

♻ ☆ Residual Stream Analysis with Multi-Layer SAEs

Tim Lawson, Lucy Farnik, Conor Houghton, Laurence Aitchison

Sparse autoencoders (SAEs) are a promising approach to interpreting the internal representations of transformer language models. However, SAEs are usually trained separately on each transformer layer, making it difficult to use them to study how information flows across layers. To solve this problem, we introduce the multi-layer SAE (MLSAE): a single SAE trained on the residual stream activation vectors from every transformer layer. Given that the residual stream is understood to preserve information across layers, we expected MLSAE latents to `switch on' at a token position and remain active at later layers. Interestingly, we find that individual latents are often active at a single layer for a given token or prompt, but this layer may differ for different tokens or prompts. We quantify these phenomena by defining a distribution over layers and considering its variance. We find that the variance of the distributions of latent activations over layers is about two orders of magnitude greater when aggregating over tokens compared with a single token. For larger underlying models, the degree to which latents are active at multiple layers increases, which is consistent with the fact that the residual stream activation vectors at adjacent layers become more similar. Finally, we relax the assumption that the residual stream basis is the same at every layer by applying pre-trained tuned-lens transformations, but our findings remain qualitatively similar. Our results represent a new approach to understanding how representations change as they flow through transformers. We release our code to train and analyze MLSAEs at https://github.com/tim-lawson/mlsae.

comment: 34 pages, 26 figures

♻ ☆ Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation

Shengjie Ma, Chengjin Xu, Xuhui Jiang, Muzhi Li, Huaren Qu, Cehao Yang, Jiaxin Mao, Jian Guo

♻ ☆ Language in Vivo vs. in Silico: Size Matters but Larger Language Models Still Do Not Comprehend Language on a Par with Humans

Vittoria Dentella, Fritz Guenther, Evelina Leivada

Understanding the limits of language is a prerequisite for Large Language Models (LLMs) to act as theories of natural language. LLM performance in some language tasks presents both quantitative and qualitative differences from that of humans, however it remains to be determined whether such differences are amenable to model size. This work investigates the critical role of model scaling, determining whether increases in size make up for such differences between humans and models. We test three LLMs from different families (Bard, 137 billion parameters; ChatGPT-3.5, 175 billion; ChatGPT-4, 1.5 trillion) on a grammaticality judgment task featuring anaphora, center embedding, comparatives, and negative polarity. N=1,200 judgments are collected and scored for accuracy, stability, and improvements in accuracy upon repeated presentation of a prompt. Results of the best performing LLM, ChatGPT-4, are compared to results of n=80 humans on the same stimuli. We find that humans are overall less accurate than ChatGPT-4 (76% vs. 80% accuracy, respectively), but that this is due to ChatGPT-4 outperforming humans only in one task condition, namely on grammatical sentences. Additionally, ChatGPT-4 wavers more than humans in its answers (12.5% vs. 9.6% likelihood of an oscillating answer, respectively). Thus, while increased model size may lead to better performance, LLMs are still not sensitive to (un)grammaticality the same way as humans are. It seems possible but unlikely that scaling alone can fix this issue. We interpret these results by comparing language learning in vivo and in silico, identifying three critical differences concerning (i) the type of evidence, (ii) the poverty of the stimulus, and (iii) the occurrence of semantic hallucinations due to impenetrable linguistic reference.

♻ ☆ SH2: Self-Highlighted Hesitation Helps You Decode More Truthfully EMNLP 2024

Jushi Kai, Tianhang Zhang, Hai Hu, Zhouhan Lin

Large language models (LLMs) demonstrate great performance in text generation. However, LLMs are still suffering from hallucinations. In this work, we propose an inference-time method, Self-Highlighted Hesitation (SH2), to help LLMs decode more truthfully. SH2 is based on a simple fact rooted in information theory that for an LLM, the tokens predicted with lower probabilities are prone to be more informative than others. Our analysis shows that the tokens assigned with lower probabilities by an LLM are more likely to be closely related to factual information, such as nouns, proper nouns, and adjectives. Therefore, we propose to ''highlight'' the factual information by selecting the tokens with the lowest probabilities and concatenating them to the original context, thus forcing the model to repeatedly read and hesitate on these tokens before generation. During decoding, we also adopt contrastive decoding to emphasize the difference in the output probabilities brought by the hesitation. Experimental results demonstrate that our SH2, requiring no additional data or models, can effectively help LLMs elicit factual knowledge and distinguish hallucinated contexts. Significant and consistent improvements are achieved by SH2 for LLaMA-7b, LLaMA2-7b and Mistral-7b on multiple hallucination tasks.

comment: EMNLP 2024 Findings

♻ ☆ CBF-LLM: Safe Control for LLM Alignment

Yuya Miyaoka, Masaki Inoue

This paper proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. The presented framework applies the safety filter, designed based on the CBF, to the output generation of the baseline LLM, i.e., the sequence of the token, with the aim of intervening in the generated text. The overall text-generation system is implemented with Llama 3 and a RoBERTa model, and the source code is available at https://github.com/Mya-Mya/CBF-LLM. The experiment demonstrates its control ability and effectiveness in reducing the number of interventions needed for user-specified alignment tasks.

♻ ☆ A Comparison of Language Modeling and Translation as Multilingual Pretraining Objectives EMNLP 2024

Zihao Li, Shaoxiong Ji, Timothee Mickus, Vincent Segonne, Jörg Tiedemann

Pretrained language models (PLMs) display impressive performances and have captured the attention of the NLP community. Establishing best practices in pretraining has, therefore, become a major focus of NLP research, especially since insights gained from monolingual English models may not necessarily apply to more complex multilingual models. One significant caveat of the current state of the art is that different works are rarely comparable: they often discuss different parameter counts, training data, and evaluation methodology. This paper proposes a comparison of multilingual pretraining objectives in a controlled methodological environment. We ensure that training data and model architectures are comparable, and discuss the downstream performances across 6 languages that we observe in probing and fine-tuning scenarios. We make two key observations: (1) the architecture dictates which pretraining objective is optimal; (2) multilingual translation is a very effective pretraining objective under the right conditions. We make our code, data, and model weights available at \texttt{\url{https://github.com/Helsinki-NLP/lm-vs-mt}}.

comment: Proceedings of EMNLP 2024

♻ ☆ What the Harm? Quantifying the Tangible Impact of Gender Bias in Machine Translation with a Human-centered Study EMNLP 2024

Beatrice Savoldi, Sara Papi, Matteo Negri, Ana Guerberof, Luisa Bentivogli

Gender bias in machine translation (MT) is recognized as an issue that can harm people and society. And yet, advancements in the field rarely involve people, the final MT users, or inform how they might be impacted by biased technologies. Current evaluations are often restricted to automatic methods, which offer an opaque estimate of what the downstream impact of gender disparities might be. We conduct an extensive human-centered study to examine if and to what extent bias in MT brings harms with tangible costs, such as quality of service gaps across women and men. To this aim, we collect behavioral data from 90 participants, who post-edited MT outputs to ensure correct gender translation. Across multiple datasets, languages, and types of users, our study shows that feminine post-editing demands significantly more technical and temporal effort, also corresponding to higher financial costs. Existing bias measurements, however, fail to reflect the found disparities. Our findings advocate for human-centered approaches that can inform the societal impact of bias.

comment: Accepted ad EMNLP 2024

♻ ☆ OffsetBias: Leveraging Debiased Data for Tuning Evaluators EMNLP2024

Junsoo Park, Seungyeon Jwa, Meiying Ren, Daeyoung Kim, Sanghyuk Choi

Employing Large Language Models (LLMs) to assess the quality of generated responses, such as prompting instruct-tuned models or fine-tuning judge models, has become a widely adopted evaluation method. It is also known that such evaluators are vulnerable to biases, such as favoring longer responses. While it is important to overcome this problem, the specifics of these biases remain under-explored. In this work, we qualitatively identify six types of biases inherent in various judge models. We propose EvalBiasBench as a meta-evaluation collection of hand-crafted test cases for each bias type. Additionally, we present de-biasing dataset construction methods and the associated preference dataset OffsetBias. Experimental results demonstrate that fine-tuning on our dataset significantly enhances the robustness of judge models against biases and improves performance across most evaluation scenarios. We release our datasets and the fine-tuned judge model to public.

comment: EMNLP2024 Findings

♻ ☆ Can Large Language Models Understand Symbolic Graphics Programs?

Zeju Qiu, Weiyang Liu, Haiwen Feng, Zhen Liu, Tim Z. Xiao, Katherine M. Collins, Joshua B. Tenenbaum, Adrian Weller, Michael J. Black, Bernhard Schölkopf

Against the backdrop of enthusiasm for large language models (LLMs), there is an urgent need to scientifically assess their capabilities and shortcomings. This is nontrivial in part because it is difficult to find tasks which the models have not encountered during training. Utilizing symbolic graphics programs, we propose a domain well-suited to test multiple spatial-semantic reasoning skills of LLMs. Popular in computer graphics, these programs procedurally generate visual data. While LLMs exhibit impressive skills in general program synthesis and analysis, symbolic graphics programs offer a new layer of evaluation: they allow us to test an LLM's ability to answer different-grained semantic-level questions of the images or 3D geometries without a vision encoder. To semantically understand the symbolic programs, LLMs would need to possess the ability to "imagine" and reason how the corresponding graphics content would look with only the symbolic description. We use this task to evaluate LLMs by creating a large benchmark for the semantic visual understanding of symbolic graphics programs, built procedurally with minimal human effort. Particular emphasis is placed on transformations of images that leave the image level semantics invariant while introducing significant changes to the underlying program. We evaluate commercial and open-source LLMs on our benchmark to assess their ability to reason about visual output of programs, finding that LLMs considered stronger at reasoning generally perform better. Lastly, we introduce a novel method to improve this ability -- Symbolic Instruction Tuning (SIT), in which the LLM is finetuned with pre-collected instruction data on symbolic graphics programs. Interestingly, we find that SIT not only improves LLM's understanding on symbolic programs, but it also improves general reasoning ability on various other benchmarks.

comment: Technical Report v2 (46 pages, 24 figures, project page: https://sgp-bench.github.io/, substantial update from v1)

♻ ☆ Jailbreak Antidote: Runtime Safety-Utility Balance via Sparse Representation Adjustment in Large Language Models

Guobin Shen, Dongcheng Zhao, Yiting Dong, Xiang He, Yi Zeng

As large language models (LLMs) become integral to various applications, ensuring both their safety and utility is paramount. Jailbreak attacks, which manipulate LLMs into generating harmful content, pose significant challenges to this balance. Existing defenses, such as prompt engineering and safety fine-tuning, often introduce computational overhead, increase inference latency, and lack runtime flexibility. Moreover, overly restrictive safety measures can degrade model utility by causing refusals of benign queries. In this paper, we introduce Jailbreak Antidote, a method that enables real-time adjustment of LLM safety preferences by manipulating a sparse subset of the model's internal states during inference. By shifting the model's hidden representations along a safety direction with varying strengths, we achieve flexible control over the safety-utility balance without additional token overhead or inference delays. Our analysis reveals that safety-related information in LLMs is sparsely distributed; adjusting approximately 5% of the internal state is as effective as modifying the entire state. Extensive experiments on nine LLMs (ranging from 2 billion to 72 billion parameters), evaluated against ten jailbreak attack methods and compared with six defense strategies, validate the effectiveness and efficiency of our approach. By directly manipulating internal states during reasoning, Jailbreak Antidote offers a lightweight, scalable solution that enhances LLM safety while preserving utility, opening new possibilities for real-time safety mechanisms in widely-deployed AI systems.

comment: 10 pages, 5 figures

♻ ☆ Towards a Universal Method for Meaningful Signal Detection

Louis Mahon

It is known that human speech and certain animal vocalizations can convey meaningful content because we can decipher the content that a given utterance does convey. This paper explores an alternative approach to determining whether a signal is meaningful, one that analyzes only the signal itself and is independent of what the conveyed meaning might be. We devise a method that takes a waveform as input and outputs a score indicating its degree of `meaningfulness`. We cluster contiguous portions of the input to minimize the total description length, and then take the length of the code of the assigned cluster labels as meaningfulness score. We evaluate our method empirically, against several baselines, and show that it is the only one to give a high score to human speech in various languages and with various speakers, a moderate score to animal vocalizations from birds and orcas, and a low score to ambient noise from various sources.

♻ ☆ Beyond Persuasion: Towards Conversational Recommender System with Credible Explanations EMNLP 2024

Peixin Qin, Chen Huang, Yang Deng, Wenqiang Lei, Tat-Seng Chua

With the aid of large language models, current conversational recommender system (CRS) has gaining strong abilities to persuade users to accept recommended items. While these CRSs are highly persuasive, they can mislead users by incorporating incredible information in their explanations, ultimately damaging the long-term trust between users and the CRS. To address this, we propose a simple yet effective method, called PC-CRS, to enhance the credibility of CRS's explanations during persuasion. It guides the explanation generation through our proposed credibility-aware persuasive strategies and then gradually refines explanations via post-hoc self-reflection. Experimental results demonstrate the efficacy of PC-CRS in promoting persuasive and credible explanations. Further analysis reveals the reason behind current methods producing incredible explanations and the potential of credible explanations to improve recommendation accuracy.

comment: Findings of EMNLP 2024. Our code is available at https://github.com/mumen798/PC-CRS

♻ ☆ ColPali: Efficient Document Retrieval with Vision Language Models

Manuel Faysse, Hugues Sibille, Tony Wu, Bilel Omrani, Gautier Viaud, Céline Hudelot, Pierre Colombo

Documents are visually rich structures that convey information through text, as well as tables, figures, page layouts, or fonts. While modern document retrieval systems exhibit strong performance on query-to-text matching, they struggle to exploit visual cues efficiently, hindering their performance on practical document retrieval applications such as Retrieval Augmented Generation. To benchmark current systems on visually rich document retrieval, we introduce the Visual Document Retrieval Benchmark ViDoRe, composed of various page-level retrieving tasks spanning multiple domains, languages, and settings. The inherent shortcomings of modern systems motivate the introduction of a new retrieval model architecture, ColPali, which leverages the document understanding capabilities of recent Vision Language Models to produce high-quality contextualized embeddings solely from images of document pages. Combined with a late interaction matching mechanism, ColPali largely outperforms modern document retrieval pipelines while being drastically faster and end-to-end trainable.

comment: Under Review

♻ ☆ Lighthouse: A User-Friendly Library for Reproducible Video Moment Retrieval and Highlight Detection EMNLP2024

Taichi Nishimura, Shota Nakada, Hokuto Munakata, Tatsuya Komatsu

We propose Lighthouse, a user-friendly library for reproducible video moment retrieval and highlight detection (MR-HD). Although researchers proposed various MR-HD approaches, the research community holds two main issues. The first is a lack of comprehensive and reproducible experiments across various methods, datasets, and video-text features. This is because no unified training and evaluation codebase covers multiple settings. The second is user-unfriendly design. Because previous works use different libraries, researchers set up individual environments. In addition, most works release only the training codes, requiring users to implement the whole inference process of MR-HD. Lighthouse addresses these issues by implementing a unified reproducible codebase that includes six models, three features, and five datasets. In addition, it provides an inference API and web demo to make these methods easily accessible for researchers and developers. Our experiments demonstrate that Lighthouse generally reproduces the reported scores in the reference papers. The code is available at https://github.com/line/lighthouse.

comment: accepted at EMNLP2024 - system demonstration track

♻ ☆ Typing to Listen at the Cocktail Party: Text-Guided Target Speaker Extraction

Xiang Hao, Jibin Wu, Jianwei Yu, Chenglin Xu, Kay Chen Tan

Humans can easily isolate a single speaker from a complex acoustic environment, a capability referred to as the "Cocktail Party Effect." However, replicating this ability has been a significant challenge in the field of target speaker extraction (TSE). Traditional TSE approaches predominantly rely on voiceprints, which raise privacy concerns and face issues related to the quality and availability of enrollment samples, as well as intra-speaker variability. To address these issues, this work introduces a novel text-guided TSE paradigm named LLM-TSE. In this paradigm, a state-of-the-art large language model, LLaMA 2, processes typed text input from users to extract semantic cues. We demonstrate that textual descriptions alone can effectively serve as cues for extraction, thus addressing privacy concerns and reducing dependency on voiceprints. Furthermore, our approach offers flexibility by allowing the user to specify the extraction or suppression of a speaker and enhances robustness against intra-speaker variability by incorporating context-dependent textual information. Experimental results show competitive performance with text-based cues alone and demonstrate the effectiveness of using text as a task selector. Additionally, they achieve a new state-of-the-art when combining text-based cues with pre-registered cues. This work represents the first integration of LLMs with TSE, potentially establishing a new benchmark in solving the cocktail party problem and expanding the scope of TSE applications by providing a versatile, privacy-conscious solution.

comment: Under review, https://github.com/haoxiangsnr/llm-tse

♻ ☆ Adaptive Contrastive Decoding in Retrieval-Augmented Generation for Handling Noisy Contexts EMNLP 2024

Youna Kim, Hyuhng Joon Kim, Cheonbok Park, Choonghyun Park, Hyunsoo Cho, Junyeob Kim, Kang Min Yoo, Sang-goo Lee, Taeuk Kim

When using large language models (LLMs) in knowledge-intensive tasks, such as open-domain question answering, external context can bridge the gap between external knowledge and the LLMs' parametric knowledge. Recent research has been developed to amplify contextual knowledge over the parametric knowledge of LLMs with contrastive decoding approaches. While these approaches could yield truthful responses when relevant context is provided, they are prone to vulnerabilities when faced with noisy contexts. We extend the scope of previous studies to encompass noisy contexts and propose adaptive contrastive decoding (ACD) to leverage contextual influence effectively. ACD demonstrates improvements in open-domain question answering tasks compared to baselines, especially in robustness by remaining undistracted by noisy contexts in retrieval-augmented generation.

comment: EMNLP 2024 Findings

♻ ☆ IndicVoices-R: Unlocking a Massive Multilingual Multi-speaker Speech Corpus for Scaling Indian TTS NeurIPS 2024

Ashwin Sankar, Srija Anand, Praveen Srinivasa Varadhan, Sherry Thomas, Mehak Singal, Shridhar Kumar, Deovrat Mehendale, Aditi Krishana, Giri Raju, Mitesh Khapra

Recent advancements in text-to-speech (TTS) synthesis show that large-scale models trained with extensive web data produce highly natural-sounding output. However, such data is scarce for Indian languages due to the lack of high-quality, manually subtitled data on platforms like LibriVox or YouTube. To address this gap, we enhance existing large-scale ASR datasets containing natural conversations collected in low-quality environments to generate high-quality TTS training data. Our pipeline leverages the cross-lingual generalization of denoising and speech enhancement models trained on English and applied to Indian languages. This results in IndicVoices-R (IV-R), the largest multilingual Indian TTS dataset derived from an ASR dataset, with 1,704 hours of high-quality speech from 10,496 speakers across 22 Indian languages. IV-R matches the quality of gold-standard TTS datasets like LJSpeech, LibriTTS, and IndicTTS. We also introduce the IV-R Benchmark, the first to assess zero-shot, few-shot, and many-shot speaker generalization capabilities of TTS models on Indian voices, ensuring diversity in age, gender, and style. We demonstrate that fine-tuning an English pre-trained model on a combined dataset of high-quality IndicTTS and our IV-R dataset results in better zero-shot speaker generalization compared to fine-tuning on the IndicTTS dataset alone. Further, our evaluation reveals limited zero-shot generalization for Indian voices in TTS models trained on prior datasets, which we improve by fine-tuning the model on our data containing diverse set of speakers across language families. We open-source all data and code, releasing the first TTS model for all 22 official Indian languages.

comment: Accepted to NeurIPS 2024 Datasets and Benchmarks track

♻ ☆ CMR Scaling Law: Predicting Critical Mixture Ratios for Continual Pre-training of Language Models EMNLP 2024

Jiawei Gu, Zacc Yang, Chuanghao Ding, Rui Zhao, Fei Tan

Large Language Models (LLMs) excel in diverse tasks but often underperform in specialized fields due to limited domain-specific or proprietary corpus. Continual pre-training (CPT) enhances LLM capabilities by imbuing new domain-specific or proprietary knowledge while replaying general corpus to prevent catastrophic forgetting. The data mixture ratio of general corpus and domain-specific corpus, however, has been chosen heuristically, leading to sub-optimal training efficiency in practice. In this context, we attempt to re-visit the scaling behavior of LLMs under the hood of CPT, and discover a power-law relationship between loss, mixture ratio, and training tokens scale. We formalize the trade-off between general and domain-specific capabilities, leading to a well-defined Critical Mixture Ratio (CMR) of general and domain data. By striking the balance, CMR maintains the model's general ability and achieves the desired domain transfer, ensuring the highest utilization of available resources. Considering the balance between efficiency and effectiveness, CMR can be regarded as the optimal mixture ratio. Through extensive experiments, we ascertain the predictability of CMR, propose CMR scaling law and have substantiated its generalization. These findings offer practical guidelines for optimizing LLM training in specialized domains, ensuring both general and domain-specific performance while efficiently managing training resources.

comment: EMNLP 2024 main conference

♻ ☆ DAPE: Data-Adaptive Positional Encoding for Length Extrapolation NeurIPS 2024

Chuanyang Zheng, Yihang Gao, Han Shi, Minbin Huang, Jingyao Li, Jing Xiong, Xiaozhe Ren, Michael Ng, Xin Jiang, Zhenguo Li, Yu Li

Positional encoding plays a crucial role in transformers, significantly impacting model performance and length generalization. Prior research has introduced absolute positional encoding (APE) and relative positional encoding (RPE) to distinguish token positions in given sequences. However, both APE and RPE remain fixed after model training regardless of input data, limiting their adaptability and flexibility. Hence, we expect that the desired positional encoding should be data-adaptive and can be dynamically adjusted with the given attention. In this paper, we propose a Data-Adaptive Positional Encoding (DAPE) method, which dynamically and semantically adjusts based on input context and learned fixed priors. Experimental validation on real-world datasets (Arxiv, Books3, and CHE) demonstrates that DAPE enhances model performances in terms of trained length and length generalization, where the improvements are statistically significant. The model visualization suggests that our model can keep both local and anti-local information. Finally, we successfully train the model on sequence length 128 and achieve better performance at evaluation sequence length 8192, compared with other static positional encoding methods, revealing the benefit of the adaptive positional encoding method.

comment: Accepted to NeurIPS 2024

♻ ☆ Progressive-Hint Prompting Improves Reasoning in Large Language Models ICML

Chuanyang Zheng, Zhengying Liu, Enze Xie, Zhenguo Li, Yu Li

The performance of Large Language Models (LLMs) in reasoning tasks depends heavily on prompt design, with Chain-of-Thought (CoT) and self-consistency being critical methods that enhance this ability. However, these methods do not fully exploit the answers generated by the LLM to guide subsequent responses. This paper proposes a new prompting method, named Progressive-Hint Prompting (PHP), that enables automatic multiple interactions between users and LLMs by using previously generated answers as hints to progressively guide toward the correct answers. PHP is orthogonal to CoT and self-consistency, making it easy to combine with state-of-the-art techniques to further improve performance. We conducted extensive and comprehensive experiments on seven benchmarks. The results show that PHP significantly improves accuracy while remaining highly efficient. For instance, with text-davinci-003, we observed a 4.2% improvement on GSM8K with greedy decoding compared to Complex CoT, and a 46.17% reduction in sample paths with self-consistency. With GPT-4 and PHP, we achieve state-of-the-art performances on SVAMP (89.1% -> 91.9%), GSM8K (92% -> 95.5%), AQuA (76.4% -> 79.9%) and MATH (50.3% -> 53.9%).

comment: Accepted to ICML AI4MATH 2024

♻ ☆ sDPO: Don't Use Your Data All at Once

Dahyun Kim, Yungi Kim, Wonho Song, Hyeonwoo Kim, Yunsu Kim, Sanghoon Kim, Chanjun Park

As development of large language models (LLM) progresses, aligning them with human preferences has become increasingly important. We propose stepwise DPO (sDPO), an extension of the recently popularized direct preference optimization (DPO) for alignment tuning. This approach involves dividing the available preference datasets and utilizing them in a stepwise manner, rather than employing it all at once. We demonstrate that this method facilitates the use of more precisely aligned reference models within the DPO training framework. Furthermore, sDPO trains the final model to be more performant, even outperforming other popular LLMs with more parameters.

♻ ☆ "I Like Sunnie More Than I Expected!": Exploring User Expectation and Perception of an Anthropomorphic LLM-based Conversational Agent for Well-Being Support

Siyi Wu, Julie Y. A. Cachia, Feixue Han, Bingsheng Yao, Tianyi Xie, Xuan Zhao, Dakuo Wang

The human-computer interaction (HCI) research community has a longstanding interest in exploring the mismatch between users' actual experiences and expectation toward new technologies, for instance, large language models (LLMs). In this study, we compared users' (N = 38) initial expectations against their post-interaction perceptions of two LLM-powered mental well-being intervention activity recommendation systems. Both systems have a built-in LLM to recommend a personalized well-being intervention activity, but one system (Sunnie) has an anthropomorphic conversational interaction design via elements such as appearance, persona, and natural conversation. Results showed that user engagement was high with both systems, and both systems exceeded users' expectations along the utility dimension, highlighting AI's potential to offer useful intervention activity recommendations. In addition, Sunnie further outperformed the non-anthropomorphic baseline system in relational warmth. These findings suggest that anthropomorphic conversational interaction design may be particularly effective in fostering warmth in mental health support contexts.

comment: In Submission

♻ ☆ MAG-SQL: Multi-Agent Generative Approach with Soft Schema Linking and Iterative Sub-SQL Refinement for Text-to-SQL

Wenxuan Xie, Gaochen Wu, Bowen Zhou

Recent In-Context Learning based methods have achieved remarkable success in Text-to-SQL task. However, there is still a large gap between the performance of these models and human performance on datasets with complex database schema and difficult questions, such as BIRD. Besides, existing work has neglected to supervise intermediate steps when solving questions iteratively with question decomposition methods, and the schema linking methods used in these works are very rudimentary. To address these issues, we propose MAG-SQL, a multi-agent generative approach with soft schema linking and iterative Sub-SQL refinement. In our framework, an entity-based method with tables' summary is used to select the columns in database, and a novel targets-conditions decomposition method is introduced to decompose those complex questions. Additionally, we build a iterative generating module which includes a Sub-SQL Generator and Sub-SQL Refiner, introducing external oversight for each step of generation. Through a series of ablation studies, the effectiveness of each agent in our framework has been demonstrated. When evaluated on the BIRD benchmark with GPT-4, MAG-SQL achieves an execution accuracy of 61.08%, compared to the baseline accuracy of 46.35% for vanilla GPT-4 and the baseline accuracy of 57.56% for MAC-SQL. Besides, our approach makes similar progress on Spider.

comment: 22 pages, 14 figures

♻ ☆ Representation Tuning

Christopher M. Ackerman

Activation engineering is becoming increasingly popular as a means of online control of large language models (LLMs). In this work, I extend the idea of active steering with vectors that represent a behavioral direction of interest to tuning those vectors directly into the model, obviating the need for online control. First, I identify activation vectors related to honesty in an open-source LLM (Llama- 2-13b-chat). Next, I demonstrate that model output can be made more or less honest by adding positive or negative multiples of these vectors to residual stream activations during generation. Then, I show that a similar effect can be achieved by fine-tuning the vectors directly into the model, by use of a dual loss function based on the cosine similarity of residual stream activations to the vectors combined with a standard token-based loss ("representation tuning"). Finally, I compare the generations in response to honesty-probing prompts from the resulting models to those from models fine-tuned with a token-based loss alone, and to those from the untuned model subjected to online steering. Overall, fine-tuning the vectors into the models using the cosine similarity plus token loss showed a stronger effect than online steering, and generalized better than using the standard loss, suggesting the potential utility of this approach as a safety measure. Code and data are available at https://github.com/cma1114/representation_tuning; tuned models are available at https://huggingface.co/collections/cackerman/ representation-tuning-66da1e5ab41cd1b824687d9f.

comment: 9 pages, 6 figures, 6 tables

♻ ☆ Revisiting Who's Harry Potter: Towards Targeted Unlearning from a Causal Intervention Perspective

Yujian Liu, Yang Zhang, Tommi Jaakkola, Shiyu Chang

This paper investigates Who's Harry Potter (WHP), a pioneering yet insufficiently understood method for LLM unlearning. We explore it in two steps. First, we introduce a new task of LLM targeted unlearning, where given an unlearning target (e.g., a person) and some unlearning documents, we aim to unlearn only the information about the target, rather than everything in the unlearning documents. We further argue that a successful unlearning should satisfy criteria such as not outputting gibberish, not fabricating facts about the unlearning target, and not releasing factual information under jailbreak attacks. Second, we construct a causal intervention framework for targeted unlearning, where the knowledge of the unlearning target is modeled as a confounder between LLM input and output, and the unlearning process as a deconfounding process. This framework justifies and extends WHP, deriving a simple unlearning algorithm that includes WHP as a special case. Experiments on existing and new datasets show that our approach, without explicitly optimizing for the aforementioned criteria, achieves competitive performance in all of them. Our code is available at https://github.com/UCSB-NLP-Chang/causal_unlearn.git.

♻ ☆ PILLOW: Enhancing Efficient Instruction Fine-tuning via Prompt Matching EMNLP 2023

Zhenting Qi, Xiaoyu Tan, Shaojie Shi, Chao Qu, Yinghui Xu, Yuan Qi

Instruction fine-tuning has conventionally been employed to adapt Large Language Models (LLMs) to a variety of tasks. Nonetheless, this technique often necessitates substantial computational resources, making it impractical for deployment by individuals or small-scale entities. Recently, Low-Rank Adaptation (LoRA) has become a promising alternative, offering high capabilities on par with full tuning with reduced resource overhead. However, attaining satisfactory performance through the fine-tuning of LoRA is a non-trivial challenge. In this paper, we propose PILLOW, which aims to improve LoRA's performance by a discrimination-based prompting method, leveraging LLMs' In-Context Learning ability. PILLOW incorporates a matching network that selects prompts from a user-defined prompt pool, concatenates the selected prompts with the user instruction as input, and performs inference using the LoRA-fine-tuned LLMs. Trained with Reinforcement Learning, PILLOW exhibits commensurate performance on various evaluation metrics compared with typical instruction fine-tuning methods, utilizing only consumer-grade GPU resources and exhibiting a large reduction in computational costs.

comment: Accepted by EMNLP 2023 (Industry Track), Oral Presentation

♻ ☆ MetaAligner: Towards Generalizable Multi-Objective Alignment of Language Models NeurIPS 2024

Kailai Yang, Zhiwei Liu, Qianqian Xie, Jimin Huang, Tianlin Zhang, Sophia Ananiadou

Recent advancements in large language models (LLMs) focus on aligning to heterogeneous human expectations and values via multi-objective preference alignment. However, existing methods are dependent on the policy model parameters, which require high-cost repetition of their alignment algorithms for each new policy model, and they cannot expand to unseen objectives due to their static alignment objectives. In this work, we propose Meta-Objective Aligner (MetaAligner), the first policy-agnostic and generalizable method for multi-objective preference alignment. MetaAligner models multi-objective alignment into three stages: (1) dynamic objectives reformulation algorithm reorganizes traditional alignment datasets to supervise the model on performing flexible alignment across different objectives; (2) conditional weak-to-strong correction paradigm aligns the weak outputs of fixed policy models to approach strong outputs with higher preferences in the corresponding alignment objectives, enabling plug-and-play inferences on any policy models, which significantly reduces training costs and facilitates alignment on close-source policy models; (3) generalizable inference method flexibly adjusts target objectives by updating their text descriptions in the prompts, facilitating generalizable alignment to unseen objectives. Experimental results show that MetaAligner achieves significant and balanced improvements in multi-objective alignments on 10 state-of-the-art policy models, and saves up to 93.63% of GPU training hours compared to previous alignment methods. The model also effectively aligns unseen objectives, marking the first step towards generalizable multi-objective preference alignment.

comment: Accepted by NeurIPS 2024 main track

♻ ☆ Spectra: A Comprehensive Study of Ternary, Quantized, and FP16 Language Models

Ayush Kaushal, Tejas Vaidhya, Arnab Kumar Mondal, Tejas Pandey, Aaryan Bhagat, Irina Rish

Rapid advancements in GPU computational power has outpaced memory capacity and bandwidth growth, creating bottlenecks in Large Language Model (LLM) inference. Post-training quantization is the leading method for addressing memory-related bottlenecks in LLM inference, but it suffers from significant performance degradation below 4-bit precision. This paper addresses these challenges by investigating the pretraining of low-bitwidth models specifically Ternary Language Models (TriLMs) as an alternative to traditional floating-point models (FloatLMs) and their post-training quantized versions (QuantLMs). We present Spectra LLM suite, the first open suite of LLMs spanning multiple bit-widths, including FloatLMs, QuantLMs, and TriLMs, ranging from 99M to 3.9B parameters trained on 300B tokens. Our comprehensive evaluation demonstrates that TriLMs offer superior scaling behavior in terms of model size (in bits). Surprisingly, at scales exceeding one billion parameters, TriLMs consistently outperform their QuantLM and FloatLM counterparts for a given bit size across various benchmarks. Notably, the 3.9B parameter TriLM matches the performance of the FloatLM 3.9B across all benchmarks, despite having fewer bits than FloatLM 830M. Overall, this research provides valuable insights into the feasibility and scalability of low-bitwidth language models, paving the way for the development of more efficient LLMs. To enhance understanding of low-bitwidth models, we are releasing 500+ intermediate checkpoints of the Spectra suite at \href{https://github.com/NolanoOrg/SpectraSuite}{https://github.com/NolanoOrg/SpectraSuite}.

comment: 42 pages, 21 figures, and 13 tables

♻ ☆ Evaluating and Safeguarding the Adversarial Robustness of Retrieval-Based In-Context Learning

Simon Yu, Jie He, Pasquale Minervini, Jeff Z. Pan

comment: COLM 2024, 30 pages, 6 figures

♻ ☆ FacLens: Transferable Probe for Foreseeing Non-Factuality in Large Language Models

Yanling Wang, Haoyang Li, Hao Zou, Jing Zhang, Xinlei He, Qi Li, Ke Xu

♻ ☆ Auto-Arena: Automating LLM Evaluations with Agent Peer Battles and Committee Discussions

Ruochen Zhao, Wenxuan Zhang, Yew Ken Chia, Weiwen Xu, Deli Zhao, Lidong Bing

As LLMs continuously evolve, there is an urgent need for a reliable evaluation method that delivers trustworthy results promptly. Currently, static benchmarks suffer from inflexibility and unreliability, leading users to prefer human voting platforms like Chatbot Arena. However, human evaluations require significant manual effort. To address this, we propose the Auto-Arena, an innovative framework that automates the entire evaluation process using LLM-powered agents. Firstly, an LLM examiner generates questions. Then, two LLM candidates engage in a multi-round peer battle based on individual questions, aiming at revealing their true performance differences. Finally, a committee of LLM judges collaboratively discusses and decides the winner, reducing bias and enhancing fairness. During the peer battles, we observe intriguing scenarios where the LLM candidates display competitive behaviors and even learn from the opponents. In our extensive experiments involving 15 recent LLMs, Auto-Arena shows a 92.14% correlation with human preferences, surpassing all previous expert-annotated benchmarks without any manual efforts. As a result, Auto-Arena offers a promising alternative to current human evaluation platforms for evaluating LLMs automatically.

♻ ☆ Evalverse: Unified and Accessible Library for Large Language Model Evaluation EMNLP 2024

Jihoo Kim, Wonho Song, Dahyun Kim, Yunsu Kim, Yungi Kim, Chanjun Park

This paper introduces Evalverse, a novel library that streamlines the evaluation of Large Language Models (LLMs) by unifying disparate evaluation tools into a single, user-friendly framework. Evalverse enables individuals with limited knowledge of artificial intelligence to easily request LLM evaluations and receive detailed reports, facilitated by an integration with communication platforms like Slack. Thus, Evalverse serves as a powerful tool for the comprehensive assessment of LLMs, offering both researchers and practitioners a centralized and easily accessible evaluation framework. Finally, we also provide a demo video for Evalverse, showcasing its capabilities and implementation in a two-minute format.

comment: Accepted to EMNLP 2024 Demo Track

♻ ☆ Corrective Retrieval Augmented Generation

Shi-Qi Yan, Jia-Chen Gu, Yun Zhu, Zhen-Hua Ling

Large language models (LLMs) inevitably exhibit hallucinations since the accuracy of generated texts cannot be secured solely by the parametric knowledge they encapsulate. Although retrieval-augmented generation (RAG) is a practicable complement to LLMs, it relies heavily on the relevance of retrieved documents, raising concerns about how the model behaves if retrieval goes wrong. To this end, we propose the Corrective Retrieval Augmented Generation (CRAG) to improve the robustness of generation. Specifically, a lightweight retrieval evaluator is designed to assess the overall quality of retrieved documents for a query, returning a confidence degree based on which different knowledge retrieval actions can be triggered. Since retrieval from static and limited corpora can only return sub-optimal documents, large-scale web searches are utilized as an extension for augmenting the retrieval results. Besides, a decompose-then-recompose algorithm is designed for retrieved documents to selectively focus on key information and filter out irrelevant information in them. CRAG is plug-and-play and can be seamlessly coupled with various RAG-based approaches. Experiments on four datasets covering short- and long-form generation tasks show that CRAG can significantly improve the performance of RAG-based approaches.

comment: Update results, add more analysis, and fix typos

♻ ☆ Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing

Zhangchen Xu, Fengqing Jiang, Luyao Niu, Yuntian Deng, Radha Poovendran, Yejin Choi, Bill Yuchen Lin

High-quality instruction data is critical for aligning large language models (LLMs). Although some models, such as Llama-3-Instruct, have open weights, their alignment data remain private, which hinders the democratization of AI. High human labor costs and a limited, predefined scope for prompting prevent existing open-source data creation methods from scaling effectively, potentially limiting the diversity and quality of public alignment datasets. Is it possible to synthesize high-quality instruction data at scale by extracting it directly from an aligned LLM? We present a self-synthesis method for generating large-scale alignment data named Magpie. Our key observation is that aligned LLMs like Llama-3-Instruct can generate a user query when we input only the left-side templates up to the position reserved for user messages, thanks to their auto-regressive nature. We use this method to prompt Llama-3-Instruct and generate 4 million instructions along with their corresponding responses. We perform a comprehensive analysis of the extracted data and select 300K high-quality instances. To compare Magpie data with other public instruction datasets, we fine-tune Llama-3-8B-Base with each dataset and evaluate the performance of the fine-tuned models. Our results indicate that in some tasks, models fine-tuned with Magpie perform comparably to the official Llama-3-8B-Instruct, despite the latter being enhanced with 10 million data points through supervised fine-tuning (SFT) and subsequent feedback learning. We also show that using Magpie solely for SFT can surpass the performance of previous public datasets utilized for both SFT and preference optimization, such as direct preference optimization with UltraFeedback. This advantage is evident on alignment benchmarks such as AlpacaEval, ArenaHard, and WildBench.

comment: Link: https://magpie-align.github.io/

♻ ☆ SpinQuant: LLM quantization with learned rotations

Zechun Liu, Changsheng Zhao, Igor Fedorov, Bilge Soran, Dhruv Choudhary, Raghuraman Krishnamoorthi, Vikas Chandra, Yuandong Tian, Tijmen Blankevoort

Post-training quantization (PTQ) techniques applied to weights, activations, and the KV cache greatly reduce memory usage, latency, and power consumption of Large Language Models (LLMs), but may lead to large quantization errors when outliers are present. Rotating activation or weight matrices helps remove outliers and benefits quantization. In this work, we identify a collection of applicable rotation parameterizations that lead to identical outputs in full-precision Transformer architectures while enhancing quantization accuracy. In addition, we find that some random rotations lead to much better quantization than others, with an up to 13 points difference in downstream zero-shot reasoning performance. As a result, we propose SpinQuant, a novel approach that incorporates learned rotation matrices for optimal quantized network accuracy. With 4-bit quantization of weight, activation, and KV-cache, SpinQuant narrows the accuracy gap on zero-shot reasoning tasks with full precision to merely 2.9 points on the LLaMA-2 7B model, surpassing LLM-QAT by 19.1 points and SmoothQuant by 25.0 points. Furthermore, SpinQuant also outperforms concurrent work QuaRot, which applies random rotations to remove outliers. In particular, for LLaMA-3 8B models that are hard to quantize, SpinQuant reduces the gap to full precision by up to 45.1% relative to QuaRot.

♻ ☆ FoodPuzzle: Developing Large Language Model Agents as Flavor Scientists

Tenghao Huang, Donghee Lee, John Sweeney, Jiatong Shi, Emily Steliotes, Matthew Lange, Jonathan May, Muhao Chen

Flavor development in the food industry is increasingly challenged by the need for rapid innovation and precise flavor profile creation. Traditional flavor research methods typically rely on iterative, subjective testing, which lacks the efficiency and scalability required for modern demands. This paper presents three contributions to address the challenges. Firstly, we define a new problem domain for scientific agents in flavor science, conceptualized as the generation of hypotheses for flavor profile sourcing and understanding. To facilitate research in this area, we introduce the FoodPuzzle, a challenging benchmark consisting of 978 food items and 1,766 flavor molecules profiles. We propose a novel Scientific Agent approach, integrating in-context learning and retrieval augmented techniques to generate grounded hypotheses in the domain of food science. Experimental results indicate that our model significantly surpasses traditional methods in flavor profile prediction tasks, demonstrating its potential to transform flavor development practices.

♻ ☆ The Generation Gap: Exploring Age Bias in the Value Systems of Large Language Models

Siyang Liu, Trish Maturi, Bowen Yi, Siqi Shen, Rada Mihalcea

We explore the alignment of values in Large Language Models (LLMs) with specific age groups, leveraging data from the World Value Survey across thirteen categories. Through a diverse set of prompts tailored to ensure response robustness, we find a general inclination of LLM values towards younger demographics, especially when compared to the US population. Although a general inclination can be observed, we also found that this inclination toward younger groups can be different across different value categories. Additionally, we explore the impact of incorporating age identity information in prompts and observe challenges in mitigating value discrepancies with different age cohorts. Our findings highlight the age bias in LLMs and provide insights for future work. Materials for our analysis are available at \url{ https://github.com/MichiganNLP/Age-Bias-In-LLMs}

comment: 4 pages

Computer Vision and Pattern Recognition 139

☆ Fine-Tuning CLIP's Last Visual Projector: A Few-Shot Cornucopia

Mohammad Fahes, Tuan-Hung Vu, Andrei Bursuc, Patrick Pérez, Raoul de Charette

We consider the problem of adapting a contrastively pretrained vision-language model like CLIP (Radford et al., 2021) for few-shot classification. The existing literature addresses this problem by learning a linear classifier of the frozen visual features, optimizing word embeddings, or learning external feature adapters. This paper introduces an alternative way for CLIP adaptation without adding 'external' parameters to optimize. We find that simply fine-tuning the last projection matrix of the vision encoder leads to strong performance compared to the existing baselines. Furthermore, we show that regularizing training with the distance between the fine-tuned and pretrained matrices adds reliability for adapting CLIP through this layer. Perhaps surprisingly, this approach, coined ProLIP, yields performances on par or better than state of the art on 11 few-shot classification benchmarks, few-shot domain generalization, cross-dataset transfer and test-time adaptation. Code will be made available at https://github.com/astra-vision/ProLIP .

comment: Preprint,under review

☆ Grounding Partially-Defined Events in Multimodal Data EMNLP

Kate Sanders, Reno Kriz, David Etter, Hannah Recknor, Alexander Martin, Cameron Carpenter, Jingyang Lin, Benjamin Van Durme

comment: Preprint; 9 pages; 2024 EMNLP Findings

☆ Brain Mapping with Dense Features: Grounding Cortical Semantic Selectivity in Natural Images With Vision Transformers

Andrew F. Luo, Jacob Yeung, Rushikesh Zawar, Shaurya Dewan, Margaret M. Henderson, Leila Wehbe, Michael J. Tarr

Advances in large-scale artificial neural networks have facilitated novel insights into the functional topology of the brain. Here, we leverage this approach to study how semantic categories are organized in the human visual cortex. To overcome the challenge presented by the co-occurrence of multiple categories in natural images, we introduce BrainSAIL (Semantic Attribution and Image Localization), a method for isolating specific neurally-activating visual concepts in images. BrainSAIL exploits semantically consistent, dense spatial features from pre-trained vision models, building upon their demonstrated ability to robustly predict neural activity. This method derives clean, spatially dense embeddings without requiring any additional training, and employs a novel denoising process that leverages the semantic consistency of images under random augmentations. By unifying the space of whole-image embeddings and dense visual features and then applying voxel-wise encoding models to these features, we enable the identification of specific subregions of each image which drive selectivity patterns in different areas of the higher visual cortex. We validate BrainSAIL on cortical regions with known category selectivity, demonstrating its ability to accurately localize and disentangle selectivity to diverse visual concepts. Next, we demonstrate BrainSAIL's ability to characterize high-level visual selectivity to scene properties and low-level visual features such as depth, luminance, and saturation, providing insights into the encoding of complex visual information. Finally, we use BrainSAIL to directly compare the feature selectivity of different brain encoding models across different regions of interest in visual cortex. Our innovative method paves the way for significant advances in mapping and decomposing high-level visual representations in the human brain.

☆ TextHawk2: A Large Vision-Language Model Excels in Bilingual OCR and Grounding with 16x Fewer Tokens

Ya-Qi Yu, Minghui Liao, Jiwen Zhang, Jihao Wu

Reading dense text and locating objects within images are fundamental abilities for Large Vision-Language Models (LVLMs) tasked with advanced jobs. Previous LVLMs, including superior proprietary models like GPT-4o, have struggled to excel in both tasks simultaneously. Moreover, previous LVLMs with fine-grained perception cost thousands of tokens per image, making them resource-intensive. We present TextHawk2, a bilingual LVLM featuring efficient fine-grained perception and demonstrating cutting-edge performance across general-purpose, OCR, and grounding tasks with 16 times fewer image tokens. Critical improvements include: (1) Token Compression: Building on the efficient architecture of its predecessor, TextHawk2 significantly reduces the number of tokens per image by 16 times, facilitating training and deployment of the TextHawk series with minimal resources. (2) Visual Encoder Reinforcement: We enhance the visual encoder through LVLM co-training, unlocking its potential for previously unseen tasks like Chinese OCR and grounding. (3) Data Diversity: We maintain a comparable scale of 100 million samples while diversifying the sources of pre-training data. We assess TextHawk2 across multiple benchmarks, where it consistently delivers superior performance and outperforms closed-source models of similar scale, such as achieving 78.4% accuracy on OCRBench, 81.4% accuracy on ChartQA, 89.6% ANLS on DocVQA, and 88.1% accuracy@0.5 on RefCOCOg-test.

☆ DART: A Diffusion-Based Autoregressive Motion Model for Real-Time Text-Driven Motion Control

Kaifeng Zhao, Gen Li, Siyu Tang

Text-conditioned human motion generation, which allows for user interaction through natural language, has become increasingly popular. Existing methods typically generate short, isolated motions based on a single input sentence. However, human motions are continuous and can extend over long periods, carrying rich semantics. Creating long, complex motions that precisely respond to streams of text descriptions, particularly in an online and real-time setting, remains a significant challenge. Furthermore, incorporating spatial constraints into text-conditioned motion generation presents additional challenges, as it requires aligning the motion semantics specified by text descriptions with geometric information, such as goal locations and 3D scene geometry. To address these limitations, we propose DART, a Diffusion-based Autoregressive motion primitive model for Real-time Text-driven motion control. Our model, DART, effectively learns a compact motion primitive space jointly conditioned on motion history and text inputs using latent diffusion models. By autoregressively generating motion primitives based on the preceding history and current text input, DART enables real-time, sequential motion generation driven by natural language descriptions. Additionally, the learned motion primitive space allows for precise spatial motion control, which we formulate either as a latent noise optimization problem or as a Markov decision process addressed through reinforcement learning. We present effective algorithms for both approaches, demonstrating our model's versatility and superior performance in various motion synthesis tasks. Experiments show our method outperforms existing baselines in motion realism, efficiency, and controllability. Video results are available on the project page: https://zkf1997.github.io/DART/.

☆ GS-VTON: Controllable 3D Virtual Try-on with Gaussian Splatting

Yukang Cao, Masoud Hadi, Liang Pan, Ziwei Liu

Diffusion-based 2D virtual try-on (VTON) techniques have recently demonstrated strong performance, while the development of 3D VTON has largely lagged behind. Despite recent advances in text-guided 3D scene editing, integrating 2D VTON into these pipelines to achieve vivid 3D VTON remains challenging. The reasons are twofold. First, text prompts cannot provide sufficient details in describing clothing. Second, 2D VTON results generated from different viewpoints of the same 3D scene lack coherence and spatial relationships, hence frequently leading to appearance inconsistencies and geometric distortions. To resolve these problems, we introduce an image-prompted 3D VTON method (dubbed GS-VTON) which, by leveraging 3D Gaussian Splatting (3DGS) as the 3D representation, enables the transfer of pre-trained knowledge from 2D VTON models to 3D while improving cross-view consistency. (1) Specifically, we propose a personalized diffusion model that utilizes low-rank adaptation (LoRA) fine-tuning to incorporate personalized information into pre-trained 2D VTON models. To achieve effective LoRA training, we introduce a reference-driven image editing approach that enables the simultaneous editing of multi-view images while ensuring consistency. (2) Furthermore, we propose a persona-aware 3DGS editing framework to facilitate effective editing while maintaining consistent cross-view appearance and high-quality 3D geometry. (3) Additionally, we have established a new 3D VTON benchmark, 3D-VTONBench, which facilitates comprehensive qualitative and quantitative 3D VTON evaluations. Through extensive experiments and comparative analyses with existing methods, the proposed \OM has demonstrated superior fidelity and advanced editing capabilities, affirming its effectiveness for 3D VTON.

comment: 21 pages, 11 figures

☆ SePPO: Semi-Policy Preference Optimization for Diffusion Alignment

Daoan Zhang, Guangchen Lan, Dong-Jun Han, Wenlin Yao, Xiaoman Pan, Hongming Zhang, Mingxiao Li, Pengcheng Chen, Yu Dong, Christopher Brinton, Jiebo Luo

Reinforcement learning from human feedback (RLHF) methods are emerging as a way to fine-tune diffusion models (DMs) for visual generation. However, commonly used on-policy strategies are limited by the generalization capability of the reward model, while off-policy approaches require large amounts of difficult-to-obtain paired human-annotated data, particularly in visual generation tasks. To address the limitations of both on- and off-policy RLHF, we propose a preference optimization method that aligns DMs with preferences without relying on reward models or paired human-annotated data. Specifically, we introduce a Semi-Policy Preference Optimization (SePPO) method. SePPO leverages previous checkpoints as reference models while using them to generate on-policy reference samples, which replace "losing images" in preference pairs. This approach allows us to optimize using only off-policy "winning images." Furthermore, we design a strategy for reference model selection that expands the exploration in the policy space. Notably, we do not simply treat reference samples as negative examples for learning. Instead, we design an anchor-based criterion to assess whether the reference samples are likely to be winning or losing images, allowing the model to selectively learn from the generated reference samples. This approach mitigates performance degradation caused by the uncertainty in reference sample quality. We validate SePPO across both text-to-image and text-to-video benchmarks. SePPO surpasses all previous approaches on the text-to-image benchmarks and also demonstrates outstanding performance on the text-to-video benchmarks. Code will be released in https://github.com/DwanZhang-AI/SePPO.

☆ LoTLIP: Improving Language-Image Pre-training for Long Text Understanding

Wei Wu, Kecheng Zheng, Shuailei Ma, Fan Lu, Yuxin Guo, Yifei Zhang, Wei Chen, Qingpei Guo, Yujun Shen, Zheng-Jun Zha

Understanding long text is of great demands in practice but beyond the reach of most language-image pre-training (LIP) models. In this work, we empirically confirm that the key reason causing such an issue is that the training images are usually paired with short captions, leaving certain tokens easily overshadowed by salient tokens. Towards this problem, our initial attempt is to relabel the data with long captions, however, directly learning with which may lead to performance degradation in understanding short text (e.g., in the image classification task). Then, after incorporating corner tokens to aggregate diverse textual information, we manage to help the model catch up to its original level of short text understanding yet greatly enhance its capability of long text understanding. We further look into whether the model can continuously benefit from longer captions and notice a clear trade-off between the performance and the efficiency. Finally, we validate the effectiveness of our approach using a self-constructed large-scale dataset, which consists of 100M long caption oriented text-image pairs. It is noteworthy that, on the task of long-text image retrieval, we beat the competitor using long captions with 11.1% improvement (i.e., from 72.62% to 83.72%). We will release the code, the model, and the new dataset to facilitate the reproducibility and further research. The project page is available at https://wuw2019.github.io/lotlip.

☆ Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents

Boyu Gou, Ruohan Wang, Boyuan Zheng, Yanan Xie, Cheng Chang, Yiheng Shu, Huan Sun, Yu Su

☆ TuneVLSeg: Prompt Tuning Benchmark for Vision-Language Segmentation Models ACCV 2024

Rabin Adhikari, Safal Thapaliya, Manish Dhakal, Bishesh Khanal

comment: Accepted at ACCV 2024 (oral presentation)

☆ DiffuseReg: Denoising Diffusion Model for Obtaining Deformation Fields in Unsupervised Deformable Image Registration MICCAI 2024

Yongtai Zhuo, Yiqing Shen

Deformable image registration aims to precisely align medical images from different modalities or times. Traditional deep learning methods, while effective, often lack interpretability, real-time observability and adjustment capacity during registration inference. Denoising diffusion models present an alternative by reformulating registration as iterative image denoising. However, existing diffusion registration approaches do not fully harness capabilities, neglecting the critical sampling phase that enables continuous observability during the inference. Hence, we introduce DiffuseReg, an innovative diffusion-based method that denoises deformation fields instead of images for improved transparency. We also propose a novel denoising network upon Swin Transformer, which better integrates moving and fixed images with diffusion time step throughout the denoising process. Furthermore, we enhance control over the denoising registration process with a novel similarity consistency regularization. Experiments on ACDC datasets demonstrate DiffuseReg outperforms existing diffusion registration methods by 1.32 in Dice score. The sampling process in DiffuseReg enables real-time output observability and adjustment unmatched by previous deep models.

comment: MICCAI 2024, W-AM-067, https://github.com/YutaZhuo/DiffuseReg

☆ SimO Loss: Anchor-Free Contrastive Loss for Fine-Grained Supervised Contrastive Learning

Taha Bouhsine, Imad El Aaroussi, Atik Faysal, Wang Huaxia

We introduce a novel anchor-free contrastive learning (AFCL) method leveraging our proposed Similarity-Orthogonality (SimO) loss. Our approach minimizes a semi-metric discriminative loss function that simultaneously optimizes two key objectives: reducing the distance and orthogonality between embeddings of similar inputs while maximizing these metrics for dissimilar inputs, facilitating more fine-grained contrastive learning. The AFCL method, powered by SimO loss, creates a fiber bundle topological structure in the embedding space, forming class-specific, internally cohesive yet orthogonal neighborhoods. We validate the efficacy of our method on the CIFAR-10 dataset, providing visualizations that demonstrate the impact of SimO loss on the embedding space. Our results illustrate the formation of distinct, orthogonal class neighborhoods, showcasing the method's ability to create well-structured embeddings that balance class separation with intra-class variability. This work opens new avenues for understanding and leveraging the geometric properties of learned representations in various machine learning tasks.

☆ The Dawn of Video Generation: Preliminary Explorations with SORA-like Models

Ailing Zeng, Yuhang Yang, Weidong Chen, Wei Liu

High-quality video generation, encompassing text-to-video (T2V), image-to-video (I2V), and video-to-video (V2V) generation, holds considerable significance in content creation to benefit anyone express their inherent creativity in new ways and world simulation to modeling and understanding the world. Models like SORA have advanced generating videos with higher resolution, more natural motion, better vision-language alignment, and increased controllability, particularly for long video sequences. These improvements have been driven by the evolution of model architectures, shifting from UNet to more scalable and parameter-rich DiT models, along with large-scale data expansion and refined training strategies. However, despite the emergence of DiT-based closed-source and open-source models, a comprehensive investigation into their capabilities and limitations remains lacking. Furthermore, the rapid development has made it challenging for recent benchmarks to fully cover SORA-like models and recognize their significant advancements. Additionally, evaluation metrics often fail to align with human preferences.

comment: project: https://ailab-cvc.github.io/VideoGen-Eval/

☆ Precise Model Benchmarking with Only a Few Observations EMNLP 2024

Riccardo Fogliato, Pratik Patil, Nil-Jana Akpinar, Mathew Monfort

comment: To appear at EMNLP 2024

☆ Organizing Unstructured Image Collections using Natural Language

Mingxuan Liu, Zhun Zhong, Jun Li, Gianni Franchi, Subhankar Roy, Elisa Ricci

Organizing unstructured visual data into semantic clusters is a key challenge in computer vision. Traditional deep clustering (DC) approaches focus on a single partition of data, while multiple clustering (MC) methods address this limitation by uncovering distinct clustering solutions. The rise of large language models (LLMs) and multimodal LLMs (MLLMs) has enhanced MC by allowing users to define clustering criteria in natural language. However, manually specifying criteria for large datasets is impractical. In this work, we introduce the task Semantic Multiple Clustering (SMC) that aims to automatically discover clustering criteria from large image collections, uncovering interpretable substructures without requiring human input. Our framework, Text Driven Semantic Multiple Clustering (TeDeSC), uses text as a proxy to concurrently reason over large image collections, discover partitioning criteria, expressed in natural language, and reveal semantic substructures. To evaluate TeDeSC, we introduce the COCO-4c and Food-4c benchmarks, each containing four grouping criteria and ground-truth annotations. We apply TeDeSC to various applications, such as discovering biases and analyzing social media image popularity, demonstrating its utility as a tool for automatically organizing image collections and revealing novel insights.

comment: Preprint. Project webpage: https://oatmealliu.github.io/smc.html

☆ Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality EMNLP 2024

Youngtaek Oh, Jae Won Cho, Dong-Jin Kim, In So Kweon, Junmo Kim

comment: EMNLP 2024 (Long, Main). Project page: https://ytaek-oh.github.io/fsc-clip

☆ Studying and Mitigating Biases in Sign Language Understanding Models

Katherine Atwell, Danielle Bragg, Malihe Alikhani

☆ Beyond FVD: Enhanced Evaluation Metrics for Video Generation Quality

Ge Ya, Luo, Gian Favero, Zhi Hao Luo, Alexia Jolicoeur-Martineau, Christopher Pal

☆ MARs: Multi-view Attention Regularizations for Patch-based Feature Recognition of Space Terrain ECCV 2024

Timothy Chase Jr, Karthik Dantu

The visual detection and tracking of surface terrain is required for spacecraft to safely land on or navigate within close proximity to celestial objects. Current approaches rely on template matching with pre-gathered patch-based features, which are expensive to obtain and a limiting factor in perceptual capability. While recent literature has focused on in-situ detection methods to enhance navigation and operational autonomy, robust description is still needed. In this work, we explore metric learning as the lightweight feature description mechanism and find that current solutions fail to address inter-class similarity and multi-view observational geometry. We attribute this to the view-unaware attention mechanism and introduce Multi-view Attention Regularizations (MARs) to constrain the channel and spatial attention across multiple feature views, regularizing the what and where of attention focus. We thoroughly analyze many modern metric learning losses with and without MARs and demonstrate improved terrain-feature recognition performance by upwards of 85%. We additionally introduce the Luna-1 dataset, consisting of Moon crater landmarks and reference navigation frames from NASA mission data to support future research in this difficult task. Luna-1 and source code are publicly available at https://droneslab.github.io/mars/.

comment: ECCV 2024. Project page available at https://droneslab.github.io/mars/

☆ VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks

Ziyan Jiang, Rui Meng, Xinyi Yang, Semih Yavuz, Yingbo Zhou, Wenhu Chen

comment: Technical Report

☆ MIBench: A Comprehensive Benchmark for Model Inversion Attack and Defense

Yixiang Qiu, Hongyao Yu, Hao Fang, Wenbo Yu, Bin Chen, Xuan Wang, Shu-Tao Xia, Ke Xu

comment: 23 pages

☆ Leveraging Multimodal Diffusion Models to Accelerate Imaging with Side Information

Timofey Efimov, Harry Dong, Megna Shah, Jeff Simmons, Sean Donegan, Yuejie Chi

Diffusion models have found phenomenal success as expressive priors for solving inverse problems, but their extension beyond natural images to more structured scientific domains remains limited. Motivated by applications in materials science, we aim to reduce the number of measurements required from an expensive imaging modality of interest, by leveraging side information from an auxiliary modality that is much cheaper to obtain. To deal with the non-differentiable and black-box nature of the forward model, we propose a framework to train a multimodal diffusion model over the joint modalities, turning inverse problems with black-box forward models into simple linear inpainting problems. Numerically, we demonstrate the feasibility of training diffusion models over materials imagery data, and show that our approach achieves superior image reconstruction by leveraging the available side information, requiring significantly less amount of data from the expensive microscopy modality.

☆ Human-Feedback Efficient Reinforcement Learning for Online Diffusion Model Finetuning

Ayano Hiranaka, Shang-Fu Chen, Chieh-Hsin Lai, Dongjun Kim, Naoki Murata, Takashi Shibuya, Wei-Hsiang Liao, Shao-Hua Sun, Yuki Mitsufuji

Controllable generation through Stable Diffusion (SD) fine-tuning aims to improve fidelity, safety, and alignment with human guidance. Existing reinforcement learning from human feedback methods usually rely on predefined heuristic reward functions or pretrained reward models built on large-scale datasets, limiting their applicability to scenarios where collecting such data is costly or difficult. To effectively and efficiently utilize human feedback, we develop a framework, HERO, which leverages online human feedback collected on the fly during model learning. Specifically, HERO features two key mechanisms: (1) Feedback-Aligned Representation Learning, an online training method that captures human feedback and provides informative learning signals for fine-tuning, and (2) Feedback-Guided Image Generation, which involves generating images from SD's refined initialization samples, enabling faster convergence towards the evaluator's intent. We demonstrate that HERO is 4x more efficient in online feedback for body part anomaly correction compared to the best existing method. Additionally, experiments show that HERO can effectively handle tasks like reasoning, counting, personalization, and reducing NSFW content with only 0.5K online feedback.

☆ Synthetic Generation of Dermatoscopic Images with GAN and Closed-Form Factorization

Rohan Reddy Mekala, Frederik Pahde, Simon Baur, Sneha Chandrashekar, Madeline Diep, Markus Wenzel, Eric L. Wisotzky, Galip Ümit Yolcu, Sebastian Lapuschkin, Jackie Ma, Peter Eisert, Mikael Lindvall, Adam Porter, Wojciech Samek

In the realm of dermatological diagnoses, where the analysis of dermatoscopic and microscopic skin lesion images is pivotal for the accurate and early detection of various medical conditions, the costs associated with creating diverse and high-quality annotated datasets have hampered the accuracy and generalizability of machine learning models. We propose an innovative unsupervised augmentation solution that harnesses Generative Adversarial Network (GAN) based models and associated techniques over their latent space to generate controlled semiautomatically-discovered semantic variations in dermatoscopic images. We created synthetic images to incorporate the semantic variations and augmented the training data with these images. With this approach, we were able to increase the performance of machine learning models and set a new benchmark amongst non-ensemble based models in skin lesion classification on the HAM10000 dataset; and used the observed analytics and generated models for detailed studies on model explainability, affirming the effectiveness of our solution.

comment: This preprint has been submitted to the Workshop on Synthetic Data for Computer Vision (SyntheticData4CV 2024 is a side event on 18th European Conference on Computer Vision 2024). This preprint has not undergone peer review or any post-submission improvements or corrections

☆ LiDAR-GS:Real-time LiDAR Re-Simulation using Gaussian Splatting

Qifeng Chen, Sheng Yang, Sicong Du, Tao Tang, Peng Chen, Yuchi Huo

LiDAR simulation plays a crucial role in closed-loop simulation for autonomous driving. Although recent advancements, such as the use of reconstructed mesh and Neural Radiance Fields (NeRF), have made progress in simulating the physical properties of LiDAR, these methods have struggled to achieve satisfactory frame rates and rendering quality. To address these limitations, we present LiDAR-GS, the first LiDAR Gaussian Splatting method, for real-time high-fidelity re-simulation of LiDAR sensor scans in public urban road scenes. The vanilla Gaussian Splatting, designed for camera models, cannot be directly applied to LiDAR re-simulation. To bridge the gap between passive camera and active LiDAR, our LiDAR-GS designs a differentiable laser beam splatting, grounded in the LiDAR range view model. This innovation allows for precise surface splatting by projecting lasers onto micro cross-sections, effectively eliminating artifacts associated with local affine approximations. Additionally, LiDAR-GS leverages Neural Gaussian Fields, which further integrate view-dependent clues, to represent key LiDAR properties that are influenced by the incident angle and external factors. Combining these practices with some essential adaptations, e.g., dynamic instances decomposition, our approach succeeds in simultaneously re-simulating depth, intensity, and ray-drop channels, achieving state-of-the-art results in both rendering frame rate and quality on publically available large scene datasets. Our source code will be made publicly available.

☆ MetaDD: Boosting Dataset Distillation with Neural Network Architecture-Invariant Generalization

Yunlong Zhao, Xiaoheng Deng, Xiu Su, Hongyan Xu, Xiuxing Li, Yijing Liu, Shan You

Dataset distillation (DD) entails creating a refined, compact distilled dataset from a large-scale dataset to facilitate efficient training. A significant challenge in DD is the dependency between the distilled dataset and the neural network (NN) architecture used. Training a different NN architecture with a distilled dataset distilled using a specific architecture often results in diminished trainning performance for other architectures. This paper introduces MetaDD, designed to enhance the generalizability of DD across various NN architectures. Specifically, MetaDD partitions distilled data into meta features (i.e., the data's common characteristics that remain consistent across different NN architectures) and heterogeneous features (i.e., the data's unique feature to each NN architecture). Then, MetaDD employs an architecture-invariant loss function for multi-architecture feature alignment, which increases meta features and reduces heterogeneous features in distilled data. As a low-memory consumption component, MetaDD can be seamlessly integrated into any DD methodology. Experimental results demonstrate that MetaDD significantly improves performance across various DD methods. On the Distilled Tiny-Imagenet with Sre2L (50 IPC), MetaDD achieves cross-architecture NN accuracy of up to 30.1\%, surpassing the second-best method (GLaD) by 1.7\%.

☆ IGroupSS-Mamba: Interval Group Spatial-Spectral Mamba for Hyperspectral Image Classification

Yan He, Bing Tu, Puzhao Jiang, Bo Liu, Jun Li, Antonio Plaza

Hyperspectral image (HSI) classification has garnered substantial attention in remote sensing fields. Recent Mamba architectures built upon the Selective State Space Models (S6) have demonstrated enormous potential in long-range sequence modeling. However, the high dimensionality of hyperspectral data and information redundancy pose challenges to the application of Mamba in HSI classification, suffering from suboptimal performance and computational efficiency. In light of this, this paper investigates a lightweight Interval Group Spatial-Spectral Mamba framework (IGroupSS-Mamba) for HSI classification, which allows for multi-directional and multi-scale global spatial-spectral information extraction in a grouping and hierarchical manner. Technically, an Interval Group S6 Mechanism (IGSM) is developed as the core component, which partitions high-dimensional features into multiple non-overlapping groups at intervals, and then integrates a unidirectional S6 for each group with a specific scanning direction to achieve non-redundant sequence modeling. Compared to conventional applying multi-directional scanning to all bands, this grouping strategy leverages the complementary strengths of different scanning directions while decreasing computational costs. To adequately capture the spatial-spectral contextual information, an Interval Group Spatial-Spectral Block (IGSSB) is introduced, in which two IGSM-based spatial and spectral operators are cascaded to characterize the global spatial-spectral relationship along the spatial and spectral dimensions, respectively. IGroupSS-Mamba is constructed as a hierarchical structure stacked by multiple IGSSB blocks, integrating a pixel aggregation-based downsampling strategy for multiscale spatial-spectral semantic learning from shallow to deep stages. Extensive experiments demonstrate that IGroupSS-Mamba outperforms the state-of-the-art methods.

☆ DreamSat: Towards a General 3D Model for Novel View Synthesis of Space Objects

Nidhi Mathihalli, Audrey Wei, Giovanni Lavezzi, Peng Mun Siew, Victor Rodriguez-Fernandez, Hodei Urrutxua, Richard Linares

Novel view synthesis (NVS) enables to generate new images of a scene or convert a set of 2D images into a comprehensive 3D model. In the context of Space Domain Awareness, since space is becoming increasingly congested, NVS can accurately map space objects and debris, improving the safety and efficiency of space operations. Similarly, in Rendezvous and Proximity Operations missions, 3D models can provide details about a target object's shape, size, and orientation, allowing for better planning and prediction of the target's behavior. In this work, we explore the generalization abilities of these reconstruction techniques, aiming to avoid the necessity of retraining for each new scene, by presenting a novel approach to 3D spacecraft reconstruction from single-view images, DreamSat, by fine-tuning the Zero123 XL, a state-of-the-art single-view reconstruction model, on a high-quality dataset of 190 high-quality spacecraft models and integrating it into the DreamGaussian framework. We demonstrate consistent improvements in reconstruction quality across multiple metrics, including Contrastive Language-Image Pretraining (CLIP) score (+0.33%), Peak Signal-to-Noise Ratio (PSNR) (+2.53%), Structural Similarity Index (SSIM) (+2.38%), and Learned Perceptual Image Patch Similarity (LPIPS) (+0.16%) on a test set of 30 previously unseen spacecraft images. Our method addresses the lack of domain-specific 3D reconstruction tools in the space industry by leveraging state-of-the-art diffusion models and 3D Gaussian splatting techniques. This approach maintains the efficiency of the DreamGaussian framework while enhancing the accuracy and detail of spacecraft reconstructions. The code for this work can be accessed on GitHub (https://github.com/ARCLab-MIT/space-nvs).

comment: Presented at the 75th International Astronautical Congress, October 2024, Milan, Italy

☆ Human-in-the-loop Reasoning For Traffic Sign Detection: Collaborative Approach Yolo With Video-llava

Mehdi Azarafza, Fatima Idrees, Ali Ehteshami Bejnordi, Charles Steinmetz, Stefan Henkler, Achim Rettberg

Traffic Sign Recognition (TSR) detection is a crucial component of autonomous vehicles. While You Only Look Once (YOLO) is a popular real-time object detection algorithm, factors like training data quality and adverse weather conditions (e.g., heavy rain) can lead to detection failures. These failures can be particularly dangerous when visual similarities between objects exist, such as mistaking a 30 km/h sign for a higher speed limit sign. This paper proposes a method that combines video analysis and reasoning, prompting with a human-in-the-loop guide large vision model to improve YOLOs accuracy in detecting road speed limit signs, especially in semi-real-world conditions. It is hypothesized that the guided prompting and reasoning abilities of Video-LLava can enhance YOLOs traffic sign detection capabilities. This hypothesis is supported by an evaluation based on human-annotated accuracy metrics within a dataset of recorded videos from the CARLA car simulator. The results demonstrate that a collaborative approach combining YOLO with Video-LLava and reasoning can effectively address challenging situations such as heavy rain and overcast conditions that hinder YOLOs detection capabilities.

comment: 10 pages, 6 figures

☆ xLSTM-FER: Enhancing Student Expression Recognition with Extended Vision Long Short-Term Memory Network APWeb

Qionghao Huang, Jili Chen

Student expression recognition has become an essential tool for assessing learning experiences and emotional states. This paper introduces xLSTM-FER, a novel architecture derived from the Extended Long Short-Term Memory (xLSTM), designed to enhance the accuracy and efficiency of expression recognition through advanced sequence processing capabilities for student facial expression recognition. xLSTM-FER processes input images by segmenting them into a series of patches and leveraging a stack of xLSTM blocks to handle these patches. xLSTM-FER can capture subtle changes in real-world students' facial expressions and improve recognition accuracy by learning spatial-temporal relationships within the sequence. Experiments on CK+, RAF-DF, and FERplus demonstrate the potential of xLSTM-FER in expression recognition tasks, showing better performance compared to state-of-the-art methods on standard datasets. The linear computational and memory complexity of xLSTM-FER make it particularly suitable for handling high-resolution images. Moreover, the design of xLSTM-FER allows for efficient processing of non-sequential inputs such as images without additional computation.

comment: The paper, consisting of 10 pages and 3 figures, has been accepted by the AIEDM Workshop at the 8th APWeb-WAIM Joint International Conference on Web and Big Data

☆ Control-oriented Clustering of Visual Latent Representation

Han Qi, Haocheng Yin, Heng Yang

☆ Improving Object Detection via Local-global Contrastive Learning BMVC 2024

Danai Triantafyllidou, Sarah Parisot, Ales Leonardis, Steven McDonagh

Visual domain gaps often impact object detection performance. Image-to-image translation can mitigate this effect, where contrastive approaches enable learning of the image-to-image mapping under unsupervised regimes. However, existing methods often fail to handle content-rich scenes with multiple object instances, which manifests in unsatisfactory detection performance. Sensitivity to such instance-level content is typically only gained through object annotations, which can be expensive to obtain. Towards addressing this issue, we present a novel image-to-image translation method that specifically targets cross-domain object detection. We formulate our approach as a contrastive learning framework with an inductive prior that optimises the appearance of object instances through spatial attention masks, implicitly delineating the scene into foreground regions associated with the target object instances and background non-object regions. Instead of relying on object annotations to explicitly account for object instances during translation, our approach learns to represent objects by contrasting local-global information. This affords investigation of an under-explored challenge: obtaining performant detection, under domain shifts, without relying on object annotations nor detector model fine-tuning. We experiment with multiple cross-domain object detection settings across three challenging benchmarks and report state-of-the-art performance. Project page: https://local-global-detection.github.io

comment: BMVC 2024 - Project page: https://local-global-detection.github.io

☆ SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification NeurIPS 2024

Benjamin Feuer, Jiawei Xu, Niv Cohen, Patrick Yubeaton, Govind Mittal, Chinmay Hegde

Data curation is the problem of how to collect and organize samples into a dataset that supports efficient learning. Despite the centrality of the task, little work has been devoted towards a large-scale, systematic comparison of various curation methods. In this work, we take steps towards a formal evaluation of data curation strategies and introduce SELECT, the first large-scale benchmark of curation strategies for image classification. In order to generate baseline methods for the SELECT benchmark, we create a new dataset, ImageNet++, which constitutes the largest superset of ImageNet-1K to date. Our dataset extends ImageNet with 5 new training-data shifts, each approximately the size of ImageNet-1K itself, and each assembled using a distinct curation strategy. We evaluate our data curation baselines in two ways: (i) using each training-data shift to train identical image classification models from scratch (ii) using the data itself to fit a pretrained self-supervised representation. Our findings show interesting trends, particularly pertaining to recent methods for data curation such as synthetic data generation and lookup based on CLIP embeddings. We show that although these strategies are highly competitive for certain tasks, the curation strategy used to assemble the original ImageNet-1K dataset remains the gold standard. We anticipate that our benchmark can illuminate the path for new methods to further reduce the gap. We release our checkpoints, code, documentation, and a link to our dataset at https://github.com/jimmyxu123/SELECT.

comment: NeurIPS 2024, Datasets and Benchmarks Track

☆ HE-Drive: Human-Like End-to-End Driving with Vision Language Models

Junming Wang, Xingyu Zhang, Zebin Xing, Songen Gu, Xiaoyang Guo, Yang Hu, Ziying Song, Qian Zhang, Xiaoxiao Long, Wei Yin

In this paper, we propose HE-Drive: the first human-like-centric end-to-end autonomous driving system to generate trajectories that are both temporally consistent and comfortable. Recent studies have shown that imitation learning-based planners and learning-based trajectory scorers can effectively generate and select accuracy trajectories that closely mimic expert demonstrations. However, such trajectory planners and scorers face the dilemma of generating temporally inconsistent and uncomfortable trajectories. To solve the above problems, Our HE-Drive first extracts key 3D spatial representations through sparse perception, which then serves as conditional inputs for a Conditional Denoising Diffusion Probabilistic Models (DDPMs)-based motion planner to generate temporal consistency multi-modal trajectories. A Vision-Language Models (VLMs)-guided trajectory scorer subsequently selects the most comfortable trajectory from these candidates to control the vehicle, ensuring human-like end-to-end driving. Experiments show that HE-Drive not only achieves state-of-the-art performance (i.e., reduces the average collision rate by 71% than VAD) and efficiency (i.e., 1.9X faster than SparseDrive) on the challenging nuScenes and OpenScene datasets but also provides the most comfortable driving experience on real-world data.For more information, visit the project website: https://jmwang0117.github.io/HE-Drive/.

☆ PhotoReg: Photometrically Registering 3D Gaussian Splatting Models

Ziwen Yuan, Tianyi Zhang, Matthew Johnson-Roberson, Weiming Zhi

Building accurate representations of the environment is critical for intelligent robots to make decisions during deployment. Advances in photorealistic environment models have enabled robots to develop hyper-realistic reconstructions, which can be used to generate images that are intuitive for human inspection. In particular, the recently introduced \ac{3DGS}, which describes the scene with up to millions of primitive ellipsoids, can be rendered in real time. \ac{3DGS} has rapidly gained prominence. However, a critical unsolved problem persists: how can we fuse multiple \ac{3DGS} into a single coherent model? Solving this problem will enable robot teams to jointly build \ac{3DGS} models of their surroundings. A key insight of this work is to leverage the {duality} between photorealistic reconstructions, which render realistic 2D images from 3D structure, and \emph{3D foundation models}, which predict 3D structure from image pairs. To this end, we develop PhotoReg, a framework to register multiple photorealistic \ac{3DGS} models with 3D foundation models. As \ac{3DGS} models are generally built from monocular camera images, they have \emph{arbitrary scale}. To resolve this, PhotoReg actively enforces scale consistency among the different \ac{3DGS} models by considering depth estimates within these models. Then, the alignment is iteratively refined with fine-grained photometric losses to produce high-quality fused \ac{3DGS} models. We rigorously evaluate PhotoReg on both standard benchmark datasets and our custom-collected datasets, including with two quadruped robots. The code is released at \url{ziweny11.github.io/photoreg}.

☆ Systematic Literature Review of Vision-Based Approaches to Outdoor Livestock Monitoring with Lessons from Wildlife Studies

Stacey D. Scott, Zayn J. Abbas, Feerass Ellid, Eli-Henry Dykhne, Muhammad Muhaiminul Islam, Weam Ayad, Kristina Kacmorova, Dan Tulpan, Minglun Gong

Precision livestock farming (PLF) aims to improve the health and welfare of livestock animals and farming outcomes through the use of advanced technologies. Computer vision, combined with recent advances in machine learning and deep learning artificial intelligence approaches, offers a possible solution to the PLF ideal of 24/7 livestock monitoring that helps facilitate early detection of animal health and welfare issues. However, a significant number of livestock species are raised in large outdoor habitats that pose technological challenges for computer vision approaches. This review provides a comprehensive overview of computer vision methods and open challenges in outdoor animal monitoring. We include research from both the livestock and wildlife fields in the review because of the similarities in appearance, behaviour, and habitat for many livestock and wildlife. We focus on large terrestrial mammals, such as cattle, horses, deer, goats, sheep, koalas, giraffes, and elephants. We use an image processing pipeline to frame our discussion and highlight the current capabilities and open technical challenges at each stage of the pipeline. The review found a clear trend towards the use of deep learning approaches for animal detection, counting, and multi-species classification. We discuss in detail the applicability of current vision-based methods to PLF contexts and promising directions for future research.

comment: 28 pages, 5 figures, 2 tables

☆ Conditional Variational Autoencoders for Probabilistic Pose Regression IROS 2024

Fereidoon Zangeneh, Leonard Bruns, Amit Dekel, Alessandro Pieropan, Patric Jensfelt

Robots rely on visual relocalization to estimate their pose from camera images when they lose track. One of the challenges in visual relocalization is repetitive structures in the operation environment of the robot. This calls for probabilistic methods that support multiple hypotheses for robot's pose. We propose such a probabilistic method to predict the posterior distribution of camera poses given an observed image. Our proposed training strategy results in a generative model of camera poses given an image, which can be used to draw samples from the pose posterior distribution. Our method is streamlined and well-founded in theory and outperforms existing methods on localization in presence of ambiguities.

comment: Accepted at IROS 2024

☆ RoWeeder: Unsupervised Weed Mapping through Crop-Row Detection ECCV 2024

Pasquale De Marinis, Rino Vessio, Giovanna Castellano

comment: Computer Vision for Plant Phenotyping and Agriculture (CVPPA) workshop at ECCV 2024

☆ Comparison of marker-less 2D image-based methods for infant pose estimation

Lennart Jahn, Sarah Flügge, Dajie Zhang, Luise Poustka, Sven Bölte, Florentin Wörgötter, Peter B Marschik, Tomas Kulvicius

There are increasing efforts to automate clinical methods for early diagnosis of developmental disorders, among them the General Movement Assessment (GMA), a video-based tool to classify infant motor functioning. Optimal pose estimation is a crucial part of the automated GMA. In this study we compare the performance of available generic- and infant-pose estimators, and the choice of viewing angle for optimal recordings, i.e., conventional diagonal view used in GMA vs. top-down view. For this study, we used 4500 annotated video-frames from 75 recordings of infant spontaneous motor functions from 4 to 26 weeks. To determine which available pose estimation method and camera angle yield the best pose estimation accuracy on infants in a GMA related setting, the distance to human annotations as well as the percentage of correct key-points (PCK) were computed and compared. The results show that the best performing generic model trained on adults, ViTPose, also performs best on infants. We see no improvement from using specialized infant-pose estimators over the generic pose estimators on our own infant dataset. However, when retraining a generic model on our data, there is a significant improvement in pose estimation accuracy. The pose estimation accuracy obtained from the top-down view is significantly better than that obtained from the diagonal view, especially for the detection of the hip key-points. The results also indicate only limited generalization capabilities of infant-pose estimators to other infant datasets, which hints that one should be careful when choosing infant pose estimators and using them on infant datasets which they were not trained on. While the standard GMA method uses a diagonal view for assessment, pose estimation accuracy significantly improves using a top-down view. This suggests that a top-down view should be included in recording setups for automated GMA research.

☆ 6DGS: Enhanced Direction-Aware Gaussian Splatting for Volumetric Rendering

Zhongpai Gao, Benjamin Planche, Meng Zheng, Anwesa Choudhuri, Terrence Chen, Ziyan Wu

Novel view synthesis has advanced significantly with the development of neural radiance fields (NeRF) and 3D Gaussian splatting (3DGS). However, achieving high quality without compromising real-time rendering remains challenging, particularly for physically-based ray tracing with view-dependent effects. Recently, N-dimensional Gaussians (N-DG) introduced a 6D spatial-angular representation to better incorporate view-dependent effects, but the Gaussian representation and control scheme are sub-optimal. In this paper, we revisit 6D Gaussians and introduce 6D Gaussian Splatting (6DGS), which enhances color and opacity representations and leverages the additional directional information in the 6D space for optimized Gaussian control. Our approach is fully compatible with the 3DGS framework and significantly improves real-time radiance field rendering by better modeling view-dependent effects and fine details. Experiments demonstrate that 6DGS significantly outperforms 3DGS and N-DG, achieving up to a 15.73 dB improvement in PSNR with a reduction of 66.5% Gaussian points compared to 3DGS.

comment: Demo Video: https://www.youtube.com/watch?v=77wN-K6Q9aM

☆ L-C4: Language-Based Video Colorization for Creative and Consistent Color

Zheng Chang, Shuchen Weng, Huan Ouyang, Yu Li, Si Li, Boxin Shi

Automatic video colorization is inherently an ill-posed problem because each monochrome frame has multiple optional color candidates. Previous exemplar-based video colorization methods restrict the user's imagination due to the elaborate retrieval process. Alternatively, conditional image colorization methods combined with post-processing algorithms still struggle to maintain temporal consistency. To address these issues, we present Language-based video Colorization for Creative and Consistent Colors (L-C4) to guide the colorization process using user-provided language descriptions. Our model is built upon a pre-trained cross-modality generative model, leveraging its comprehensive language understanding and robust color representation abilities. We introduce the cross-modality pre-fusion module to generate instance-aware text embeddings, enabling the application of creative colors. Additionally, we propose temporally deformable attention to prevent flickering or color shifts, and cross-clip fusion to maintain long-term color consistency. Extensive experimental results demonstrate that L-C4 outperforms relevant methods, achieving semantically accurate colors, unrestricted creative correspondence, and temporally robust consistency.

☆ Revealing Directions for Text-guided 3D Face Editing

Zhuo Chen, Yichao Yan, Sehngqi Liu, Yuhao Cheng, Weiming Zhao, Lincheng Li, Mengxiao Bi, Xiaokang Yang

3D face editing is a significant task in multimedia, aimed at the manipulation of 3D face models across various control signals. The success of 3D-aware GAN provides expressive 3D models learned from 2D single-view images only, encouraging researchers to discover semantic editing directions in its latent space. However, previous methods face challenges in balancing quality, efficiency, and generalization. To solve the problem, we explore the possibility of introducing the strength of diffusion model into 3D-aware GANs. In this paper, we present Face Clan, a fast and text-general approach for generating and manipulating 3D faces based on arbitrary attribute descriptions. To achieve disentangled editing, we propose to diffuse on the latent space under a pair of opposite prompts to estimate the mask indicating the region of interest on latent codes. Based on the mask, we then apply denoising to the masked latent codes to reveal the editing direction. Our method offers a precisely controllable manipulation method, allowing users to intuitively customize regions of interest with the text description. Experiments demonstrate the effectiveness and generalization of our Face Clan for various pre-trained GANs. It offers an intuitive and wide application for text-guided face editing that contributes to the landscape of multimedia content creation.

☆ On Efficient Variants of Segment Anything Model: A Survey

Xiaorui Sun, Jun Liu, Heng Tao Shen, Xiaofeng Zhu, Ping Hu

The Segment Anything Model (SAM) is a foundational model for image segmentation tasks, known for its strong generalization across diverse applications. However, its impressive performance comes with significant computational and resource demands, making it challenging to deploy in resource-limited environments such as mobile devices. To address this, a variety of SAM variants have been proposed to enhance efficiency without sacrificing accuracy. This survey provides the first comprehensive review of these efficient SAM variants. We begin by exploring the motivations driving this research. We then present core techniques used in SAM and model acceleration. This is followed by an in-depth analysis of various acceleration strategies, categorized by approach. Finally, we offer a unified and extensive evaluation of these methods, assessing their efficiency and accuracy on representative benchmarks, and providing a clear comparison of their overall performance.

comment: Report in progress

☆ Real-time Ship Recognition and Georeferencing for the Improvement of Maritime Situational Awareness

Borja Carrillo Perez

In an era where maritime infrastructures are crucial, advanced situational awareness solutions are increasingly important. The use of optical camera systems can allow real-time usage of maritime footage. This thesis presents an investigation into leveraging deep learning and computer vision to advance real-time ship recognition and georeferencing for the improvement of maritime situational awareness. A novel dataset, ShipSG, is introduced, containing 3,505 images and 11,625 ship masks with corresponding class and geographic position. After an exploration of state-of-the-art, a custom real-time segmentation architecture, ScatYOLOv8+CBAM, is designed for the NVIDIA Jetson AGX Xavier embedded system. This architecture adds the 2D scattering transform and attention mechanisms to YOLOv8, achieving an mAP of 75.46% and an 25.3 ms per frame, outperforming state-of-the-art methods by over 5%. To improve small and distant ship recognition in high-resolution images on embedded systems, an enhanced slicing mechanism is introduced, improving mAP by 8% to 11%. Additionally, a georeferencing method is proposed, achieving positioning errors of 18 m for ships up to 400 m away and 44 m for ships between 400 m and 1200 m. The findings are also applied in real-world scenarios, such as the detection of abnormal ship behaviour, camera integrity assessment and 3D reconstruction. The approach of this thesis outperforms existing methods and provides a framework for integrating recognized and georeferenced ships into real-time systems, enhancing operational effectiveness and decision-making for maritime stakeholders. This thesis contributes to the maritime computer vision field by establishing a benchmark for ship segmentation and georeferencing research, demonstrating the viability of deep-learning-based recognition and georeferencing methods for real-time maritime monitoring.

☆ Next state prediction gives rise to entangled, yet compositional representations of objects

Tankred Saanum, Luca M. Schulze Buschoff, Peter Dayan, Eric Schulz

Compositional representations are thought to enable humans to generalize across combinatorially vast state spaces. Models with learnable object slots, which encode information about objects in separate latent codes, have shown promise for this type of generalization but rely on strong architectural priors. Models with distributed representations, on the other hand, use overlapping, potentially entangled neural codes, and their ability to support compositional generalization remains underexplored. In this paper we examine whether distributed models can develop linearly separable representations of objects, like slotted models, through unsupervised training on videos of object interactions. We show that, surprisingly, models with distributed representations often match or outperform models with object slots in downstream prediction tasks. Furthermore, we find that linearly separable object representations can emerge without object-centric priors, with auxiliary objectives like next-state prediction playing a key role. Finally, we observe that distributed models' object representations are never fully disentangled, even if they are linearly separable: Multiple objects can be encoded through partially overlapping neural populations while still being highly separable with a linear classifier. We hypothesize that maintaining partially shared codes enables distributed models to better compress object dynamics, potentially enhancing generalization.

☆ PRFusion: Toward Effective and Robust Multi-Modal Place Recognition with Image and Point Cloud Fusion

Sijie Wang, Qiyu Kang, Rui She, Kai Zhao, Yang Song, Wee Peng Tay

Place recognition plays a crucial role in the fields of robotics and computer vision, finding applications in areas such as autonomous driving, mapping, and localization. Place recognition identifies a place using query sensor data and a known database. One of the main challenges is to develop a model that can deliver accurate results while being robust to environmental variations. We propose two multi-modal place recognition models, namely PRFusion and PRFusion++. PRFusion utilizes global fusion with manifold metric attention, enabling effective interaction between features without requiring camera-LiDAR extrinsic calibrations. In contrast, PRFusion++ assumes the availability of extrinsic calibrations and leverages pixel-point correspondences to enhance feature learning on local windows. Additionally, both models incorporate neural diffusion layers, which enable reliable operation even in challenging environments. We verify the state-of-the-art performance of both models on three large-scale benchmarks. Notably, they outperform existing models by a substantial margin of +3.0 AR@1 on the demanding Boreas dataset. Furthermore, we conduct ablation studies to validate the effectiveness of our proposed methods. The codes are available at: https://github.com/sijieaaa/PRFusion

comment: accepted by IEEE TITS 2024

☆ OmniBooth: Learning Latent Control for Image Synthesis with Multi-modal Instruction

Leheng Li, Weichao Qiu, Xu Yan, Jing He, Kaiqiang Zhou, Yingjie Cai, Qing Lian, Bingbing Liu, Ying-Cong Chen

We present OmniBooth, an image generation framework that enables spatial control with instance-level multi-modal customization. For all instances, the multimodal instruction can be described through text prompts or image references. Given a set of user-defined masks and associated text or image guidance, our objective is to generate an image, where multiple objects are positioned at specified coordinates and their attributes are precisely aligned with the corresponding guidance. This approach significantly expands the scope of text-to-image generation, and elevates it to a more versatile and practical dimension in controllability. In this paper, our core contribution lies in the proposed latent control signals, a high-dimensional spatial feature that provides a unified representation to integrate the spatial, textual, and image conditions seamlessly. The text condition extends ControlNet to provide instance-level open-vocabulary generation. The image condition further enables fine-grained control with personalized identity. In practice, our method empowers users with more flexibility in controllable generation, as users can choose multi-modal conditions from text or images as needed. Furthermore, thorough experiments demonstrate our enhanced performance in image synthesis fidelity and alignment across different tasks and datasets. Project page: https://len-li.github.io/omnibooth-web/

☆ Art2Mus: Bridging Visual Arts and Music through Cross-Modal Generation ECCV 2024

Ivan Rinaldi, Nicola Fanelli, Giovanna Castellano, Gennaro Vessio

Artificial Intelligence and generative models have revolutionized music creation, with many models leveraging textual or visual prompts for guidance. However, existing image-to-music models are limited to simple images, lacking the capability to generate music from complex digitized artworks. To address this gap, we introduce $\mathcal{A}\textit{rt2}\mathcal{M}\textit{us}$, a novel model designed to create music from digitized artworks or text inputs. $\mathcal{A}\textit{rt2}\mathcal{M}\textit{us}$ extends the AudioLDM~2 architecture, a text-to-audio model, and employs our newly curated datasets, created via ImageBind, which pair digitized artworks with music. Experimental results demonstrate that $\mathcal{A}\textit{rt2}\mathcal{M}\textit{us}$ can generate music that resonates with the input stimuli. These findings suggest promising applications in multimedia art, interactive installations, and AI-driven creative tools.

comment: Presented at the AI for Visual Arts (AI4VA) workshop at ECCV 2024

☆ D-PoSE: Depth as an Intermediate Representation for 3D Human Pose and Shape Estimation

Nikolaos Vasilikopoulos, Drosakis Drosakis, Antonis Argyros

We present D-PoSE (Depth as an Intermediate Representation for 3D Human Pose and Shape Estimation), a one-stage method that estimates human pose and SMPL-X shape parameters from a single RGB image. Recent works use larger models with transformer backbones and decoders to improve the accuracy in human pose and shape (HPS) benchmarks. D-PoSE proposes a vision based approach that uses the estimated human depth-maps as an intermediate representation for HPS and leverages training with synthetic data and the ground-truth depth-maps provided with them for depth supervision during training. Although trained on synthetic datasets, D-PoSE achieves state-of-the-art performance on the real-world benchmark datasets, EMDB and 3DPW. Despite its simple lightweight design and the CNN backbone, it outperforms ViT-based models that have a number of parameters that is larger by almost an order of magnitude. D-PoSE code is available at: https://github.com/nvasilik/D-PoSE

☆ Patch is Enough: Naturalistic Adversarial Patch against Vision-Language Pre-training Models

Dehong Kong, Siyuan Liang, Xiaopeng Zhu, Yuansheng Zhong, Wenqi Ren

Visual language pre-training (VLP) models have demonstrated significant success across various domains, yet they remain vulnerable to adversarial attacks. Addressing these adversarial vulnerabilities is crucial for enhancing security in multimodal learning. Traditionally, adversarial methods targeting VLP models involve simultaneously perturbing images and text. However, this approach faces notable challenges: first, adversarial perturbations often fail to translate effectively into real-world scenarios; second, direct modifications to the text are conspicuously visible. To overcome these limitations, we propose a novel strategy that exclusively employs image patches for attacks, thus preserving the integrity of the original text. Our method leverages prior knowledge from diffusion models to enhance the authenticity and naturalness of the perturbations. Moreover, to optimize patch placement and improve the efficacy of our attacks, we utilize the cross-attention mechanism, which encapsulates intermodal interactions by generating attention maps to guide strategic patch placements. Comprehensive experiments conducted in a white-box setting for image-to-text scenarios reveal that our proposed method significantly outperforms existing techniques, achieving a 100% attack success rate. Additionally, it demonstrates commendable performance in transfer tasks involving text-to-image configurations.

comment: accepted by Visual Intelligence

☆ Improved detection of discarded fish species through BoxAL active learning

Maria Sokolova, Pieter M. Blok, Angelo Mencarelli, Arjan Vroegop, Aloysius van Helmond, Gert Kootstra

In recent years, powerful data-driven deep-learning techniques have been developed and applied for automated catch registration. However, these methods are dependent on the labelled data, which is time-consuming, labour-intensive, expensive to collect and need expert knowledge. In this study, we present an active learning technique, named BoxAL, which includes estimation of epistemic certainty of the Faster R-CNN object-detection model. The method allows selecting the most uncertain training images from an unlabeled pool, which are then used to train the object-detection model. To evaluate the method, we used an open-source image dataset obtained with a dedicated image-acquisition system developed for commercial trawlers targeting demersal species. We demonstrated, that our approach allows reaching the same object-detection performance as with the random sampling using 400 fewer labelled images. Besides, mean AP score was significantly higher at the last training iteration with 1100 training images, specifically, 39.0±1.6 and 34.8±1.8 for certainty-based sampling and random sampling, respectively. Additionally, we showed that epistemic certainty is a suitable method to sample images that the current iteration of the model cannot deal with yet. Our study additionally showed that the sampled new data is more valuable for training than the remaining unlabeled data. Our software is available on https://github.com/pieterblok/boxal.

☆ TeX-NeRF: Neural Radiance Fields from Pseudo-TeX Vision

Chonghao Zhong, Chao Xu

Neural radiance fields (NeRF) has gained significant attention for its exceptional visual effects. However, most existing NeRF methods reconstruct 3D scenes from RGB images captured by visible light cameras. In practical scenarios like darkness, low light, or bad weather, visible light cameras become ineffective. Therefore, we propose TeX-NeRF, a 3D reconstruction method using only infrared images, which introduces the object material emissivity as a priori, preprocesses the infrared images using Pseudo-TeX vision, and maps the temperatures (T), emissivities (e), and textures (X) of the scene into the saturation (S), hue (H), and value (V) channels of the HSV color space, respectively. Novel view synthesis using the processed images has yielded excellent results. Additionally, we introduce 3D-TeX Datasets, the first dataset comprising infrared images and their corresponding Pseudo-TeX vision images. Experiments demonstrate that our method not only matches the quality of scene reconstruction achieved with high-quality RGB images but also provides accurate temperature estimations for objects in the scene.

☆ Art Forgery Detection using Kolmogorov Arnold and Convolutional Neural Networks ECCV 2024

Sandro Boccuzzo, Deborah Desirée Meyer, Ludovica Schaerf

Art authentication has historically established itself as a task requiring profound connoisseurship of one particular artist. Nevertheless, famous art forgers such as Wolfgang Beltracchi were able to deceive dozens of art experts. In recent years Artificial Intelligence algorithms have been successfully applied to various image processing tasks. In this work, we leverage the growing improvements in AI to present an art authentication framework for the identification of the forger Wolfgang Beltracchi. Differently from existing literature on AI-aided art authentication, we focus on a specialized model of a forger, rather than an artist, flipping the approach of traditional AI methods. We use a carefully compiled dataset of known artists forged by Beltracchi and a set of known works by the forger to train a multiclass image classification model based on EfficientNet. We compare the results with Kolmogorov Arnold Networks (KAN) which, to the best of our knowledge, have never been tested in the art domain. The results show a general agreement between the different models' predictions on artworks flagged as forgeries, which are then closely studied using visual analysis.

comment: Accepted to ECCV 2024 workshop AI4VA, oral presentation

☆ Causal Context Adjustment Loss for Learned Image Compression NeurIPS 2024

Minghao Han, Shiyin Jiang, Shengxi Li, Xin Deng, Mai Xu, Ce Zhu, Shuhang Gu

In recent years, learned image compression (LIC) technologies have surpassed conventional methods notably in terms of rate-distortion (RD) performance. Most present learned techniques are VAE-based with an autoregressive entropy model, which obviously promotes the RD performance by utilizing the decoded causal context. However, extant methods are highly dependent on the fixed hand-crafted causal context. The question of how to guide the auto-encoder to generate a more effective causal context benefit for the autoregressive entropy models is worth exploring. In this paper, we make the first attempt in investigating the way to explicitly adjust the causal context with our proposed Causal Context Adjustment loss (CCA-loss). By imposing the CCA-loss, we enable the neural network to spontaneously adjust important information into the early stage of the autoregressive entropy model. Furthermore, as transformer technology develops remarkably, variants of which have been adopted by many state-of-the-art (SOTA) LIC techniques. The existing computing devices have not adapted the calculation of the attention mechanism well, which leads to a burden on computation quantity and inference latency. To overcome it, we establish a convolutional neural network (CNN) image compression model and adopt the unevenly channel-wise grouped strategy for high efficiency. Ultimately, the proposed CNN-based LIC network trained with our Causal Context Adjustment loss attains a great trade-off between inference latency and rate-distortion performance.

comment: Accepted to NeurIPS 2024

☆ PostEdit: Posterior Sampling for Efficient Zero-Shot Image Editing

Feng Tian, Yixuan Li, Yichao Yan, Shanyan Guan, Yanhao Ge, Xiaokang Yang

In the field of image editing, three core challenges persist: controllability, background preservation, and efficiency. Inversion-based methods rely on time-consuming optimization to preserve the features of the initial images, which results in low efficiency due to the requirement for extensive network inference. Conversely, inversion-free methods lack theoretical support for background similarity, as they circumvent the issue of maintaining initial features to achieve efficiency. As a consequence, none of these methods can achieve both high efficiency and background consistency. To tackle the challenges and the aforementioned disadvantages, we introduce PostEdit, a method that incorporates a posterior scheme to govern the diffusion sampling process. Specifically, a corresponding measurement term related to both the initial features and Langevin dynamics is introduced to optimize the estimated image generated by the given target prompt. Extensive experimental results indicate that the proposed PostEdit achieves state-of-the-art editing performance while accurately preserving unedited regions. Furthermore, the method is both inversion- and training-free, necessitating approximately 1.5 seconds and 18 GB of GPU memory to generate high-quality results.

☆ A Simple Image Segmentation Framework via In-Context Examples NeurIPS

Yang Liu, Chenchen Jing, Hengtao Li, Muzhi Zhu, Hao Chen, Xinlong Wang, Chunhua Shen

comment: Accepted to Proc. Conference on Neural Information Processing Systems (NeurIPS) 2024. Webpage: https://github.com/aim-uofa/SINE

☆ Multimodal Fusion Strategies for Mapping Biophysical Landscape Features ECCV 2024

Lucia Gordon, Nico Lang, Catherine Ressijac, Andrew Davies

Multimodal aerial data are used to monitor natural systems, and machine learning can significantly accelerate the classification of landscape features within such imagery to benefit ecology and conservation. It remains under-explored, however, how these multiple modalities ought to be fused in a deep learning model. As a step towards filling this gap, we study three strategies (Early fusion, Late fusion, and Mixture of Experts) for fusing thermal, RGB, and LiDAR imagery using a dataset of spatially-aligned orthomosaics in these three modalities. In particular, we aim to map three ecologically-relevant biophysical landscape features in African savanna ecosystems: rhino middens, termite mounds, and water. The three fusion strategies differ in whether the modalities are fused early or late, and if late, whether the model learns fixed weights per modality for each class or generates weights for each class adaptively, based on the input. Overall, the three methods have similar macro-averaged performance with Late fusion achieving an AUC of 0.698, but their per-class performance varies strongly, with Early fusion achieving the best recall for middens and water and Mixture of Experts achieving the best recall for mounds.

comment: 9 pages, 4 figures, ECCV 2024 Workshop in CV for Ecology

☆ CAT: Concept-level backdoor ATtacks for Concept Bottleneck Models

Songning Lai, Jiayu Yang, Yu Huang, Lijie Hu, Tianlang Xue, Zhangyi Hu, Jiaxu Li, Haicheng Liao, Yutao Yue

Despite the transformative impact of deep learning across multiple domains, the inherent opacity of these models has driven the development of Explainable Artificial Intelligence (XAI). Among these efforts, Concept Bottleneck Models (CBMs) have emerged as a key approach to improve interpretability by leveraging high-level semantic information. However, CBMs, like other machine learning models, are susceptible to security threats, particularly backdoor attacks, which can covertly manipulate model behaviors. Understanding that the community has not yet studied the concept level backdoor attack of CBM, because of "Better the devil you know than the devil you don't know.", we introduce CAT (Concept-level Backdoor ATtacks), a methodology that leverages the conceptual representations within CBMs to embed triggers during training, enabling controlled manipulation of model predictions at inference time. An enhanced attack pattern, CAT+, incorporates a correlation function to systematically select the most effective and stealthy concept triggers, thereby optimizing the attack's impact. Our comprehensive evaluation framework assesses both the attack success rate and stealthiness, demonstrating that CAT and CAT+ maintain high performance on clean data while achieving significant targeted effects on backdoored datasets. This work underscores the potential security risks associated with CBMs and provides a robust testing methodology for future security assessments.

☆ Resource-Efficient Multiview Perception: Integrating Semantic Masking with Masked Autoencoders

Kosta Dakic, Kanchana Thilakarathna, Rodrigo N. Calheiros, Teng Joon Lim

Multiview systems have become a key technology in modern computer vision, offering advanced capabilities in scene understanding and analysis. However, these systems face critical challenges in bandwidth limitations and computational constraints, particularly for resource-limited camera nodes like drones. This paper presents a novel approach for communication-efficient distributed multiview detection and tracking using masked autoencoders (MAEs). We introduce a semantic-guided masking strategy that leverages pre-trained segmentation models and a tunable power function to prioritize informative image regions. This approach, combined with an MAE, reduces communication overhead while preserving essential visual information. We evaluate our method on both virtual and real-world multiview datasets, demonstrating comparable performance in terms of detection and tracking performance metrics compared to state-of-the-art techniques, even at high masking ratios. Our selective masking algorithm outperforms random masking, maintaining higher accuracy and precision as the masking ratio increases. Furthermore, our approach achieves a significant reduction in transmission data volume compared to baseline methods, thereby balancing multiview tracking performance with communication efficiency.

comment: 10 pages, conference

☆ Learning Efficient and Effective Trajectories for Differential Equation-based Image Restoration

Zhiyu Zhu, Jinhui Hou, Hui Liu, Huanqiang Zeng, Junhui Hou

The differential equation-based image restoration approach aims to establish learnable trajectories connecting high-quality images to a tractable distribution, e.g., low-quality images or a Gaussian distribution. In this paper, we reformulate the trajectory optimization of this kind of method, focusing on enhancing both reconstruction quality and efficiency. Initially, we navigate effective restoration paths through a reinforcement learning process, gradually steering potential trajectories toward the most precise options. Additionally, to mitigate the considerable computational burden associated with iterative sampling, we propose cost-aware trajectory distillation to streamline complex paths into several manageable steps with adaptable sizes. Moreover, we fine-tune a foundational diffusion model (FLUX) with 12B parameters by using our algorithms, producing a unified framework for handling 7 kinds of image restoration tasks. Extensive experiments showcase the significant superiority of the proposed method, achieving a maximum PSNR improvement of 2.1 dB over state-of-the-art methods, while also greatly enhancing visual perceptual quality. Project page: \url{https://zhu-zhiyu.github.io/FLUX-IR/}.

☆ FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

Haokun Chen, Hang Li, Yao Zhang, Gengyuan Zhang, Jinhe Bi, Philip Torr, Jindong Gu, Denis Krompass, Volker Tresp

One-Shot Federated Learning (OSFL), a special decentralized machine learning paradigm, has recently gained significant attention. OSFL requires only a single round of client data or model upload, which reduces communication costs and mitigates privacy threats compared to traditional FL. Despite these promising prospects, existing methods face challenges due to client data heterogeneity and limited data quantity when applied to real-world OSFL systems. Recently, Latent Diffusion Models (LDM) have shown remarkable advancements in synthesizing high-quality images through pretraining on large-scale datasets, thereby presenting a potential solution to overcome these issues. However, directly applying pretrained LDM to heterogeneous OSFL results in significant distribution shifts in synthetic data, leading to performance degradation in classification models trained on such data. This issue is particularly pronounced in rare domains, such as medical imaging, which are underrepresented in LDM's pretraining data. To address this challenge, we propose Federated Bi-Level Personalization (FedBiP), which personalizes the pretrained LDM at both instance-level and concept-level. Hereby, FedBiP synthesizes images following the client's local data distribution without compromising the privacy regulations. FedBiP is also the first approach to simultaneously address feature space heterogeneity and client data scarcity in OSFL. Our method is validated through extensive experiments on three OSFL benchmarks with feature space heterogeneity, as well as on challenging medical and satellite image datasets with label heterogeneity. The results demonstrate the effectiveness of FedBiP, which substantially outperforms other OSFL methods.

☆ Building Damage Assessment in Conflict Zones: A Deep Learning Approach Using Geospatial Sub-Meter Resolution Data

Matteo Risso, Alessia Goffi, Beatrice Alessandra Motetti, Alessio Burrello, Jean Baptiste Bove, Enrico Macii, Massimo Poncino, Daniele Jahier Pagliari, Giuseppe Maffeis

Very High Resolution (VHR) geospatial image analysis is crucial for humanitarian assistance in both natural and anthropogenic crises, as it allows to rapidly identify the most critical areas that need support. Nonetheless, manually inspecting large areas is time-consuming and requires domain expertise. Thanks to their accuracy, generalization capabilities, and highly parallelizable workload, Deep Neural Networks (DNNs) provide an excellent way to automate this task. Nevertheless, there is a scarcity of VHR data pertaining to conflict situations, and consequently, of studies on the effectiveness of DNNs in those scenarios. Motivated by this, our work extensively studies the applicability of a collection of state-of-the-art Convolutional Neural Networks (CNNs) originally developed for natural disasters damage assessment in a war scenario. To this end, we build an annotated dataset with pre- and post-conflict images of the Ukrainian city of Mariupol. We then explore the transferability of the CNN models in both zero-shot and learning scenarios, demonstrating their potential and limitations. To the best of our knowledge, this is the first study to use sub-meter resolution imagery to assess building damage in combat zones.

comment: This paper has been accepted for publication in the Sixth IEEE International Conference on Image Processing Applications and Systems 2024 copyright IEEE

☆ Improving Image Clustering with Artifacts Attenuation via Inference-Time Attention Engineering ACCV 2024

Kazumoto Nakamura, Yuji Nozawa, Yu-Chieh Lin, Kengo Nakata, Youyang Ng

The goal of this paper is to improve the performance of pretrained Vision Transformer (ViT) models, particularly DINOv2, in image clustering task without requiring re-training or fine-tuning. As model size increases, high-norm artifacts anomaly appears in the patches of multi-head attention. We observe that this anomaly leads to reduced accuracy in zero-shot image clustering. These artifacts are characterized by disproportionately large values in the attention map compared to other patch tokens. To address these artifacts, we propose an approach called Inference-Time Attention Engineering (ITAE), which manipulates attention function during inference. Specifically, we identify the artifacts by investigating one of the Query-Key-Value (QKV) patches in the multi-head attention and attenuate their corresponding attention values inside the pretrained models. ITAE shows improved clustering accuracy on multiple datasets by exhibiting more expressive features in latent space. Our findings highlight the potential of ITAE as a practical solution for reducing artifacts in pretrained ViT models and improving model performance in clustering tasks without the need for re-training or fine-tuning.

comment: Accepted to ACCV 2024

☆ Transforming Color: A Novel Image Colorization Method

Hamza Shafiq, Bumshik Lee

This paper introduces a novel method for image colorization that utilizes a color transformer and generative adversarial networks (GANs) to address the challenge of generating visually appealing colorized images. Conventional approaches often struggle with capturing long-range dependencies and producing realistic colorizations. The proposed method integrates a transformer architecture to capture global information and a GAN framework to improve visual quality. In this study, a color encoder that utilizes a random normal distribution to generate color features is applied. These features are then integrated with grayscale image features to enhance the overall representation of the images. Our method demonstrates superior performance compared with existing approaches by utilizing the capacity of the transformer, which can capture long-range dependencies and generate a realistic colorization of the GAN. Experimental results show that the proposed network significantly outperforms other state-of-the-art colorization techniques, highlighting its potential for image colorization. This research opens new possibilities for precise and visually compelling image colorization in domains such as digital restoration and historical image analysis.

☆ Analysis of Hybrid Compositions in Animation Film with Weakly Supervised Learning ECCV

Mónica Apellaniz Portos, Roberto Labadie-Tamayo, Claudius Stemmler, Erwin Feyersinger, Andreas Babic, Franziska Bruckner, Vrääth Öhner, Matthias Zeppelzauer

We present an approach for the analysis of hybrid visual compositions in animation in the domain of ephemeral film. We combine ideas from semi-supervised and weakly supervised learning to train a model that can segment hybrid compositions without requiring pre-labeled segmentation masks. We evaluate our approach on a set of ephemeral films from 13 film archives. Results demonstrate that the proposed learning strategy yields a performance close to a fully supervised baseline. On a qualitative level the performed analysis provides interesting insights on hybrid compositions in animation film.

comment: Vision for Art (VISART VII) Workshop at the European Conference of Computer Vision (ECCV)

☆ Mitigating Modality Prior-Induced Hallucinations in Multimodal Large Language Models via Deciphering Attention Causality

Guanyu Zhou, Yibo Yan, Xin Zou, Kun Wang, Aiwei Liu, Xuming Hu

Multimodal Large Language Models (MLLMs) have emerged as a central focus in both industry and academia, but often suffer from biases introduced by visual and language priors, which can lead to multimodal hallucination. These biases arise from the visual encoder and the Large Language Model (LLM) backbone, affecting the attention mechanism responsible for aligning multimodal inputs. Existing decoding-based mitigation methods focus on statistical correlations and overlook the causal relationships between attention mechanisms and model output, limiting their effectiveness in addressing these biases. To tackle this issue, we propose a causal inference framework termed CausalMM that applies structural causal modeling to MLLMs, treating modality priors as a confounder between attention mechanisms and output. Specifically, by employing backdoor adjustment and counterfactual reasoning at both the visual and language attention levels, our method mitigates the negative effects of modality priors and enhances the alignment of MLLM's inputs and outputs, with a maximum score improvement of 65.3% on 6 VLind-Bench indicators and 164 points on MME Benchmark compared to conventional methods. Extensive experiments validate the effectiveness of our approach while being a plug-and-play solution. Our code is available at: https://github.com/The-Martyr/CausalMM

☆ MM-R$^3$: On (In-)Consistency of Multi-modal Large Language Models (MLLMs)

Shih-Han Chou, Shivam Chandhok, James J. Little, Leonid Sigal

With the advent of Large Language Models (LLMs) and Multimodal (Visio-lingual) LLMs, a flurry of research has emerged, analyzing the performance of such models across a diverse array of tasks. While most studies focus on evaluating the capabilities of state-of-the-art (SoTA) MLLM models through task accuracy (e.g., Visual Question Answering, grounding) across various datasets, our work explores the related but complementary aspect of consistency - the ability of an MLLM model to produce semantically similar or identical responses to semantically similar queries. We note that consistency is a fundamental prerequisite (necessary but not sufficient condition) for robustness and trust in MLLMs. Humans, in particular, are known to be highly consistent (even if not always accurate) in their responses, and consistency is inherently expected from AI systems. Armed with this perspective, we propose the MM-R$^3$ benchmark, which analyses the performance in terms of consistency and accuracy in SoTA MLLMs with three tasks: Question Rephrasing, Image Restyling, and Context Reasoning. Our analysis reveals that consistency does not always align with accuracy, indicating that models with higher accuracy are not necessarily more consistent, and vice versa. Furthermore, we propose a simple yet effective mitigation strategy in the form of an adapter module trained to minimize inconsistency across prompts. With our proposed strategy, we are able to achieve absolute improvements of 5.7% and 12.5%, on average on widely used MLLMs such as BLIP-2 and LLaVa 1.5M in terms of consistency over their existing counterparts.

☆ WTCL-Dehaze: Rethinking Real-world Image Dehazing via Wavelet Transform and Contrastive Learning

Divine Joseph Appiah, Donghai Guan, Abdul Nasser Kasule, Mingqiang Wei

Images captured in hazy outdoor conditions often suffer from colour distortion, low contrast, and loss of detail, which impair high-level vision tasks. Single image dehazing is essential for applications such as autonomous driving and surveillance, with the aim of restoring image clarity. In this work, we propose WTCL-Dehaze an enhanced semi-supervised dehazing network that integrates Contrastive Loss and Discrete Wavelet Transform (DWT). We incorporate contrastive regularization to enhance feature representation by contrasting hazy and clear image pairs. Additionally, we utilize DWT for multi-scale feature extraction, effectively capturing high-frequency details and global structures. Our approach leverages both labelled and unlabelled data to mitigate the domain gap and improve generalization. The model is trained on a combination of synthetic and real-world datasets, ensuring robust performance across different scenarios. Extensive experiments demonstrate that our proposed algorithm achieves superior performance and improved robustness compared to state-of-the-art single image dehazing methods on both benchmark datasets and real-world images.

comment: 15 pages,4 figures

☆ Intriguing Properties of Large Language and Vision Models

Young-Jun Lee, Byungsoo Ko, Han-Gyu Kim, Yechan Hwang, Ho-Jin Choi

comment: Code is available in https://github.com/passing2961/IP-LLVM

☆ LLaVA Needs More Knowledge: Retrieval Augmented Natural Language Generation with Knowledge Graph for Explaining Thoracic Pathologies

Ameer Hamza, Abdullah, Yong Hyun Ahn, Sungyoung Lee, Seong Tae Kim

Generating Natural Language Explanations (NLEs) for model predictions on medical images, particularly those depicting thoracic pathologies, remains a critical and challenging task. Existing methodologies often struggle due to general models' insufficient domain-specific medical knowledge and privacy concerns associated with retrieval-based augmentation techniques. To address these issues, we propose a novel Vision-Language framework augmented with a Knowledge Graph (KG)-based datastore, which enhances the model's understanding by incorporating additional domain-specific medical knowledge essential for generating accurate and informative NLEs. Our framework employs a KG-based retrieval mechanism that not only improves the precision of the generated explanations but also preserves data privacy by avoiding direct data retrieval. The KG datastore is designed as a plug-and-play module, allowing for seamless integration with various model architectures. We introduce and evaluate three distinct frameworks within this paradigm: KG-LLaVA, which integrates the pre-trained LLaVA model with KG-RAG; Med-XPT, a custom framework combining MedCLIP, a transformer-based projector, and GPT-2; and Bio-LLaVA, which adapts LLaVA by incorporating the Bio-ViT-L vision model. These frameworks are validated on the MIMIC-NLE dataset, where they achieve state-of-the-art results, underscoring the effectiveness of KG augmentation in generating high-quality NLEs for thoracic pathologies.

☆ Diffusion Models in 3D Vision: A Survey

Zhen Wang, Dongyuan Li, Renhe Jiang

In recent years, 3D vision has become a crucial field within computer vision, powering a wide range of applications such as autonomous driving, robotics, augmented reality (AR), and medical imaging. This field relies on the accurate perception, understanding, and reconstruction of 3D scenes from 2D data sources like images and videos. Diffusion models, originally designed for 2D generative tasks, offer the potential for more flexible, probabilistic approaches that can better capture the variability and uncertainty present in real-world 3D data. However, traditional methods often struggle with efficiency and scalability. In this paper, we review the state-of-the-art approaches that leverage diffusion models for 3D visual tasks, including but not limited to 3D object generation, shape completion, point cloud reconstruction, and scene understanding. We provide an in-depth discussion of the underlying mathematical principles of diffusion models, outlining their forward and reverse processes, as well as the various architectural advancements that enable these models to work with 3D datasets. We also discuss the key challenges in applying diffusion models to 3D vision, such as handling occlusions and varying point densities, and the computational demands of high-dimensional data. Finally, we discuss potential solutions, including improving computational efficiency, enhancing multimodal fusion, and exploring the use of large-scale pretraining for better generalization across 3D tasks. This paper serves as a foundation for future exploration and development in this rapidly evolving field.

☆ TLDR: Token-Level Detective Reward Model for Large Vision Language Models

Deqing Fu, Tong Xiao, Rui Wang, Wang Zhu, Pengchuan Zhang, Guan Pang, Robin Jia, Lawrence Chen

comment: Work done at Meta

☆ PredFormer: Transformers Are Effective Spatial-Temporal Predictive Learners

Yujin Tang, Lu Qi, Fei Xie, Xiangtai Li, Chao Ma, Ming-Hsuan Yang

Spatiotemporal predictive learning methods generally fall into two categories: recurrent-based approaches, which face challenges in parallelization and performance, and recurrent-free methods, which employ convolutional neural networks (CNNs) as encoder-decoder architectures. These methods benefit from strong inductive biases but often at the expense of scalability and generalization. This paper proposes PredFormer, a pure transformer-based framework for spatiotemporal predictive learning. Motivated by the Vision Transformers (ViT) design, PredFormer leverages carefully designed Gated Transformer blocks, following a comprehensive analysis of 3D attention mechanisms, including full-, factorized-, and interleaved- spatial-temporal attention. With its recurrent-free, transformer-based design, PredFormer is both simple and efficient, significantly outperforming previous methods by large margins. Extensive experiments on synthetic and real-world datasets demonstrate that PredFormer achieves state-of-the-art performance. On Moving MNIST, PredFormer achieves a 51.3% reduction in MSE relative to SimVP. For TaxiBJ, the model decreases MSE by 33.1% and boosts FPS from 533 to 2364. Additionally, on WeatherBench, it reduces MSE by 11.1% while enhancing FPS from 196 to 404. These performance gains in both accuracy and efficiency demonstrate PredFormer's potential for real-world applications. The source code will be released at https://github.com/yyyujintang/PredFormer.

comment: 15 pages, 7 figures

☆ ACDC: Autoregressive Coherent Multimodal Generation using Diffusion Correction

Hyungjin Chung, Dohun Lee, Jong Chul Ye

Autoregressive models (ARMs) and diffusion models (DMs) represent two leading paradigms in generative modeling, each excelling in distinct areas: ARMs in global context modeling and long-sequence generation, and DMs in generating high-quality local contexts, especially for continuous data such as images and short videos. However, ARMs often suffer from exponential error accumulation over long sequences, leading to physically implausible results, while DMs are limited by their local context generation capabilities. In this work, we introduce Autoregressive Coherent multimodal generation with Diffusion Correction (ACDC), a zero-shot approach that combines the strengths of both ARMs and DMs at the inference stage without the need for additional fine-tuning. ACDC leverages ARMs for global context generation and memory-conditioned DMs for local correction, ensuring high-quality outputs by correcting artifacts in generated multimodal tokens. In particular, we propose a memory module based on large language models (LLMs) that dynamically adjusts the conditioning texts for the DMs, preserving crucial global context information. Our experiments on multimodal tasks, including coherent multi-frame story generation and autoregressive video generation, demonstrate that ACDC effectively mitigates the accumulation of errors and significantly enhances the quality of generated outputs, achieving superior performance while remaining agnostic to specific ARM and DM architectures. Project page: https://acdc2025.github.io/

comment: 25 pages, 10 figures. Project page: https://acdc2025.github.io/

☆ H-SIREN: Improving implicit neural representations with hyperbolic periodic functions

Rui Gao, Rajeev K. Jaiman

Implicit neural representations (INR) have been recently adopted in various applications ranging from computer vision tasks to physics simulations by solving partial differential equations. Among existing INR-based works, multi-layer perceptrons with sinusoidal activation functions find widespread applications and are also frequently treated as a baseline for the development of better activation functions for INR applications. Recent investigations claim that the use of sinusoidal activation functions could be sub-optimal due to their limited supported frequency set as well as their tendency to generate over-smoothed solutions. We provide a simple solution to mitigate such an issue by changing the activation function at the first layer from $\sin(x)$ to $\sin(\sinh(2x))$. We demonstrate H-SIREN in various computer vision and fluid flow problems, where it surpasses the performance of several state-of-the-art INRs.

☆ Low-Rank Continual Pyramid Vision Transformer: Incrementally Segment Whole-Body Organs in CT with Light-Weighted Adaptation MICCAI 2024

Vince Zhu, Zhanghexuan Ji, Dazhou Guo, Puyang Wang, Yingda Xia, Le Lu, Xianghua Ye, Wei Zhu, Dakai Jin

Deep segmentation networks achieve high performance when trained on specific datasets. However, in clinical practice, it is often desirable that pretrained segmentation models can be dynamically extended to enable segmenting new organs without access to previous training datasets or without training from scratch. This would ensure a much more efficient model development and deployment paradigm accounting for the patient privacy and data storage issues. This clinically preferred process can be viewed as a continual semantic segmentation (CSS) problem. Previous CSS works would either experience catastrophic forgetting or lead to unaffordable memory costs as models expand. In this work, we propose a new continual whole-body organ segmentation model with light-weighted low-rank adaptation (LoRA). We first train and freeze a pyramid vision transformer (PVT) base segmentation model on the initial task, then continually add light-weighted trainable LoRA parameters to the frozen model for each new learning task. Through a holistically exploration of the architecture modification, we identify three most important layers (i.e., patch-embedding, multi-head attention and feed forward layers) that are critical in adapting to the new segmentation tasks, while retaining the majority of the pretrained parameters fixed. Our proposed model continually segments new organs without catastrophic forgetting and meanwhile maintaining a low parameter increasing rate. Continually trained and tested on four datasets covering different body parts of a total of 121 organs, results show that our model achieves high segmentation accuracy, closely reaching the PVT and nnUNet upper bounds, and significantly outperforms other regularization-based CSS methods. When comparing to the leading architecture-based CSS method, our model has a substantial lower parameter increasing rate while achieving comparable performance.

comment: Accepted by Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024

☆ On the Adversarial Risk of Test Time Adaptation: An Investigation into Realistic Test-Time Data Poisoning

Yongyi Su, Yushu Li, Nanqing Liu, Kui Jia, Xulei Yang, Chuan-Sheng Foo, Xun Xu

Test-time adaptation (TTA) updates the model weights during the inference stage using testing data to enhance generalization. However, this practice exposes TTA to adversarial risks. Existing studies have shown that when TTA is updated with crafted adversarial test samples, also known as test-time poisoned data, the performance on benign samples can deteriorate. Nonetheless, the perceived adversarial risk may be overstated if the poisoned data is generated under overly strong assumptions. In this work, we first review realistic assumptions for test-time data poisoning, including white-box versus grey-box attacks, access to benign data, attack budget, and more. We then propose an effective and realistic attack method that better produces poisoned samples without access to benign samples, and derive an effective in-distribution attack objective. We also design two TTA-aware attack objectives. Our benchmarks of existing attack methods reveal that the TTA methods are more robust than previously believed. In addition, we analyze effective defense strategies to help develop adversarially robust TTA methods.

comment: 19 pages, 4 figures, 8 tables

☆ Next Best Sense: Guiding Vision and Touch with FisherRF for 3D Gaussian Splatting

Matthew Strong, Boshu Lei, Aiden Swann, Wen Jiang, Kostas Daniilidis, Monroe Kennedy III

We propose a framework for active next best view and touch selection for robotic manipulators using 3D Gaussian Splatting (3DGS). 3DGS is emerging as a useful explicit 3D scene representation for robotics, as it has the ability to represent scenes in a both photorealistic and geometrically accurate manner. However, in real-world, online robotic scenes where the number of views is limited given efficiency requirements, random view selection for 3DGS becomes impractical as views are often overlapping and redundant. We address this issue by proposing an end-to-end online training and active view selection pipeline, which enhances the performance of 3DGS in few-view robotics settings. We first elevate the performance of few-shot 3DGS with a novel semantic depth alignment method using Segment Anything Model 2 (SAM2) that we supplement with Pearson depth and surface normal loss to improve color and depth reconstruction of real-world scenes. We then extend FisherRF, a next-best-view selection method for 3DGS, to select views and touch poses based on depth uncertainty. We perform online view selection on a real robot system during live 3DGS training. We motivate our improvements to few-shot GS scenes, and extend depth-based FisherRF to them, where we demonstrate both qualitative and quantitative improvements on challenging robot scenes. For more information, please see our project page at https://armlabstanford.github.io/next-best-sense.

☆ CAR: Controllable Autoregressive Modeling for Visual Generation

Ziyu Yao, Jialin Li, Yifeng Zhou, Yong Liu, Xi Jiang, Chengjie Wang, Feng Zheng, Yuexian Zou, Lei Li

Controllable generation, which enables fine-grained control over generated outputs, has emerged as a critical focus in visual generative models. Currently, there are two primary technical approaches in visual generation: diffusion models and autoregressive models. Diffusion models, as exemplified by ControlNet and T2I-Adapter, offer advanced control mechanisms, whereas autoregressive models, despite showcasing impressive generative quality and scalability, remain underexplored in terms of controllability and flexibility. In this study, we introduce Controllable AutoRegressive Modeling (CAR), a novel, plug-and-play framework that integrates conditional control into multi-scale latent variable modeling, enabling efficient control generation within a pre-trained visual autoregressive model. CAR progressively refines and captures control representations, which are injected into each autoregressive step of the pre-trained model to guide the generation process. Our approach demonstrates excellent controllability across various types of conditions and delivers higher image quality compared to previous methods. Additionally, CAR achieves robust generalization with significantly fewer training resources compared to those required for pre-training the model. To the best of our knowledge, we are the first to propose a control framework for pre-trained autoregressive visual generation models.

comment: Code available at: https://github.com/MiracleDance/CAR

☆ ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models

Ziyue Wang, Chi Chen, Fuwen Luo, Yurui Dong, Yuanchi Zhang, Yuzhuang Xu, Xiaolong Wang, Peng Li, Yang Liu

Active perception, a crucial human capability, involves setting a goal based on the current understanding of the environment and performing actions to achieve that goal. Despite significant efforts in evaluating Multimodal Large Language Models (MLLMs), active perception has been largely overlooked. To address this gap, we propose a novel benchmark named ActiView to evaluate active perception in MLLMs. Since comprehensively assessing active perception is challenging, we focus on a specialized form of Visual Question Answering (VQA) that eases the evaluation yet challenging for existing MLLMs. Given an image, we restrict the perceptual field of a model, requiring it to actively zoom or shift its perceptual field based on reasoning to answer the question successfully. We conduct extensive evaluation over 27 models, including proprietary and open-source models, and observe that the ability to read and comprehend multiple images simultaneously plays a significant role in enabling active perception. Results reveal a significant gap in the active perception capability of MLLMs, indicating that this area deserves more attention. We hope that our benchmark could help develop methods for MLLMs to understand multimodal inputs in more natural and holistic ways.

♻ ☆ Frame-Voyager: Learning to Query Frames for Video Large Language Models

Sicheng Yu, Chengkai Jin, Huanyu Wang, Zhenghao Chen, Sheng Jin, Zhongrong Zuo, Xiaolei Xu, Zhenbang Sun, Bingni Zhang, Jiawei Wu, Hao Zhang, Qianru Sun

comment: 19 pages, 10 figures

♻ ☆ Machine Learning for Asymptomatic Ratoon Stunting Disease Detection With Freely Available Satellite Based Multispectral Imaging

Ethan Kane Waters, Carla Chia-ming Chen, Mostafa Rahimi Azghadi

Disease detection in sugarcane, particularly the identification of asymptomatic infectious diseases such as Ratoon Stunting Disease (RSD), is critical for effective crop management. This study employed various machine learning techniques to detect the presence of RSD in different sugarcane varieties, using vegetation indices derived from freely available satellite-based spectral data. Our results show that the Support Vector Machine with a Radial Basis Function Kernel (SVM-RBF) was the most effective algorithm, achieving classification accuracy between 85.64% and 96.55%, depending on the variety. Gradient Boosting and Random Forest also demonstrated high performance achieving accuracy between 83.33% to 96.55%, while Logistic Regression and Quadratic Discriminant Analysis showed variable results across different varieties. The inclusion of sugarcane variety and vegetation indices was important in the detection of RSD. This agreed with what was identified in the current literature. Our study highlights the potential of satellite-based remote sensing as a cost-effective and efficient method for large-scale sugarcane disease detection alternative to traditional manual laboratory testing methods.

comment: 13 pages, 1 figure and 3 tables (main text), 1 figure and 2 tables (appendices). Submitted to "Computers and Electronics in Agriculture"

♻ ☆ CPFD: Confidence-aware Privileged Feature Distillation for Short Video Classification CIKM 2024

Jinghao Shi, Xiang Shen, Kaili Zhao, Xuedong Wang, Vera Wen, Zixuan Wang, Yifan Wu, Zhixin Zhang

Dense features, customized for different business scenarios, are essential in short video classification. However, their complexity, specific adaptation requirements, and high computational costs make them resource-intensive and less accessible during online inference. Consequently, these dense features are categorized as `Privileged Dense Features'.Meanwhile, end-to-end multi-modal models have shown promising results in numerous computer vision tasks. In industrial applications, prioritizing end-to-end multi-modal features, can enhance efficiency but often leads to the loss of valuable information from historical privileged dense features. To integrate both features while maintaining efficiency and manageable resource costs, we present Confidence-aware Privileged Feature Distillation (CPFD), which empowers features of an end-to-end multi-modal model by adaptively distilling privileged features during training. Unlike existing privileged feature distillation (PFD) methods, which apply uniform weights to all instances during distillation, potentially causing unstable performance across different business scenarios and a notable performance gap between teacher model (Dense Feature enhanced multimodal-model DF-X-VLM) and student model (multimodal-model only X-VLM), our CPFD leverages confidence scores derived from the teacher model to adaptively mitigate the performance variance with the student model. We conducted extensive offline experiments on five diverse tasks demonstrating that CPFD improves the video classification F1 score by 6.76% compared with end-to-end multimodal-model (X-VLM) and by 2.31% with vanilla PFD on-average. And it reduces the performance gap by 84.6% and achieves results comparable to teacher model DF-X-VLM. The effectiveness of CPFD is further substantiated by online experiments, and our framework has been deployed in production systems for over a dozen models.

comment: Camera ready for CIKM 2024

♻ ☆ mDPO: Conditional Preference Optimization for Multimodal Large Language Models EMNLP 2024

Fei Wang, Wenxuan Zhou, James Y. Huang, Nan Xu, Sheng Zhang, Hoifung Poon, Muhao Chen

comment: Accepted to EMNLP 2024 Main Conference. Project website: https://feiwang96.github.io/mDPO

♻ ☆ Generative Parameter-Efficient Fine-Tuning

Chinmay Savadikar, Xi Song, Tianfu Wu

We present Generative Parameter-Efficient Fine-Tuning (GIFT) for adapting pretrained Transformer backbones on downstream tasks. GIFT learns to generate the fine-tuned weights for a layer directly from its pretrained weights. The GIFT network is parameterized in a minimally-simple way by two linear layers (without bias terms), and is shared by different pretrained layers selected for fine-tuning (e.g., the Query layers), which result in significantly fewer trainable parameters compared to the layer-specific methods like Low-Rank Adapter (LoRA). We also show this formulation bridges parameter-efficient fine-tuning and representation fine-tuning. We perform comprehensive experiments on natural language tasks (commonsense and arithmetic reasoning, instruction tuning, and sequence classification) and computer vision tasks (fine-grained classification). We obtain the best performance and parameter efficiency among baselines on commonsense and arithmetic reasoning, and instruction following using the Llama family of models and on visual recognition benchmarks using Vision Transformers. Notably, compared to LoRA, we obtain 5.7% absolute increase in average accuracy with 14 times reduction of parameters on Commonsense170k using Llama-3 (8B), and 5.4% absolute increase in the win rate with 4 times reduction of parameters using Llama-2 (7B) during instruction tuning. Our GIFT also obtains a slightly higher win rate on instruction tuning than GPT 3.5 (Turbo 1106).

comment: Project page and code: https://savadikarc.github.io/gift

♻ ☆ 3D-free meets 3D priors: Novel View Synthesis from a Single Image with Pretrained Diffusion Guidance

Taewon Kang, Divya Kothandaraman, Dinesh Manocha, Ming C. Lin

comment: 13 pages, 12 figures, v2: analysis studies and more results added

♻ ☆ Finding Visual Task Vectors

Alberto Hojel, Yutong Bai, Trevor Darrell, Amir Globerson, Amir Bar

Visual Prompting is a technique for teaching models to perform a visual task via in-context examples, without any additional training. In this work, we analyze the activations of MAE-VQGAN, a recent Visual Prompting model, and find task vectors, activations that encode task-specific information. Equipped with this insight, we demonstrate that it is possible to identify the task vectors and use them to guide the network towards performing different tasks without providing any input-output examples. To find task vectors, we compute the average intermediate activations per task and use the REINFORCE algorithm to search for the subset of task vectors. The resulting task vectors guide the model towards performing a task better than the original model without the need for input-output examples.

comment: https://github.com/alhojel/visual_task_vectors

♻ ☆ A Narrative Review of Image Processing Techniques Related to Prostate Ultrasound

Haiqiao Wang, Hong Wu, Zhuoyuan Wang, Peiyan Yue, Dong Ni, Pheng-Ann Heng, Yi Wang

Prostate cancer (PCa) poses a significant threat to men's health, with early diagnosis being crucial for improving prognosis and reducing mortality rates. Transrectal ultrasound (TRUS) plays a vital role in the diagnosis and image-guided intervention of PCa.To facilitate physicians with more accurate and efficient computer-assisted diagnosis and interventions, many image processing algorithms in TRUS have been proposed and achieved state-of-the-art performance in several tasks, including prostate gland segmentation, prostate image registration, PCa classification and detection, and interventional needle detection. The rapid development of these algorithms over the past two decades necessitates a comprehensive summary. In consequence, this survey provides a \textcolor{blue}{narrative } analysis of this field, outlining the evolution of image processing methods in the context of TRUS image analysis and meanwhile highlighting their relevant contributions. Furthermore, this survey discusses current challenges and suggests future research directions to possibly advance this field further.

comment: Accepted by Ultrasound in Medicine & Biology

♻ ☆ MetaMetrics: Calibrating Metrics For Generation Tasks Using Human Preferences

Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, Derry Tanti Wijaya

comment: Preprint

♻ ☆ CYCLO: Cyclic Graph Transformer Approach to Multi-Object Relationship Modeling in Aerial Videos NeurIPS 2024

Trong-Thuan Nguyen, Pha Nguyen, Xin Li, Jackson Cothren, Alper Yilmaz, Khoa Luu

Video scene graph generation (VidSGG) has emerged as a transformative approach to capturing and interpreting the intricate relationships among objects and their temporal dynamics in video sequences. In this paper, we introduce the new AeroEye dataset that focuses on multi-object relationship modeling in aerial videos. Our AeroEye dataset features various drone scenes and includes a visually comprehensive and precise collection of predicates that capture the intricate relationships and spatial arrangements among objects. To this end, we propose the novel Cyclic Graph Transformer (CYCLO) approach that allows the model to capture both direct and long-range temporal dependencies by continuously updating the history of interactions in a circular manner. The proposed approach also allows one to handle sequences with inherent cyclical patterns and process object relationships in the correct sequential order. Therefore, it can effectively capture periodic and overlapping relationships while minimizing information loss. The extensive experiments on the AeroEye dataset demonstrate the effectiveness of the proposed CYCLO model, demonstrating its potential to perform scene understanding on drone videos. Finally, the CYCLO method consistently achieves State-of-the-Art (SOTA) results on two in-the-wild scene graph generation benchmarks, i.e., PVSG and ASPIRe.

comment: Accepted to NeurIPS 2024

♻ ☆ GMAI-MMBench: A Comprehensive Multimodal Evaluation Benchmark Towards General Medical AI

Pengcheng Chen, Jin Ye, Guoan Wang, Yanjun Li, Zhongying Deng, Wei Li, Tianbin Li, Haodong Duan, Ziyan Huang, Yanzhou Su, Benyou Wang, Shaoting Zhang, Bin Fu, Jianfei Cai, Bohan Zhuang, Eric J Seibel, Junjun He, Yu Qiao

Large Vision-Language Models (LVLMs) are capable of handling diverse data types such as imaging, text, and physiological signals, and can be applied in various fields. In the medical field, LVLMs have a high potential to offer substantial assistance for diagnosis and treatment. Before that, it is crucial to develop benchmarks to evaluate LVLMs' effectiveness in various medical applications. Current benchmarks are often built upon specific academic literature, mainly focusing on a single domain, and lacking varying perceptual granularities. Thus, they face specific challenges, including limited clinical relevance, incomplete evaluations, and insufficient guidance for interactive LVLMs. To address these limitations, we developed the GMAI-MMBench, the most comprehensive general medical AI benchmark with well-categorized data structure and multi-perceptual granularity to date. It is constructed from 284 datasets across 38 medical image modalities, 18 clinical-related tasks, 18 departments, and 4 perceptual granularities in a Visual Question Answering (VQA) format. Additionally, we implemented a lexical tree structure that allows users to customize evaluation tasks, accommodating various assessment needs and substantially supporting medical AI research and applications. We evaluated 50 LVLMs, and the results show that even the advanced GPT-4o only achieves an accuracy of 53.96%, indicating significant room for improvement. Moreover, we identified five key insufficiencies in current cutting-edge LVLMs that need to be addressed to advance the development of better medical applications. We believe that GMAI-MMBench will stimulate the community to build the next generation of LVLMs toward GMAI.

comment: GitHub: https://github.com/uni-medical/GMAI-MMBench; Hugging face: https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench

♻ ☆ Robust Multimodal Learning with Missing Modalities via Parameter-Efficient Adaptation

Md Kaykobad Reza, Ashley Prater-Bennette, M. Salman Asif

Multimodal learning seeks to utilize data from multiple sources to improve the overall performance of downstream tasks. It is desirable for redundancies in the data to make multimodal systems robust to missing or corrupted observations in some correlated modalities. However, we observe that the performance of several existing multimodal networks significantly deteriorates if one or multiple modalities are absent at test time. To enable robustness to missing modalities, we propose a simple and parameter-efficient adaptation procedure for pretrained multimodal networks. In particular, we exploit modulation of intermediate features to compensate for the missing modalities. We demonstrate that such adaptation can partially bridge performance drop due to missing modalities and outperform independent, dedicated networks trained for the available modality combinations in some cases. The proposed adaptation requires extremely small number of parameters (e.g., fewer than 1% of the total parameters) and applicable to a wide range of modality combinations and tasks. We conduct a series of experiments to highlight the missing modality robustness of our proposed method on five different multimodal tasks across seven datasets. Our proposed method demonstrates versatility across various tasks and datasets, and outperforms existing methods for robust multimodal learning with missing modalities.

comment: Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 28 pages, 6 figures, 17 tables

♻ ☆ NoSENSE: Learned unrolled cardiac MRI reconstruction without explicit sensitivity maps MICCAI

Felix Frederik Zimmermann, Andreas Kofler

We present a novel learned image reconstruction method for accelerated cardiac MRI with multiple receiver coils based on deep convolutional neural networks (CNNs) and algorithm unrolling. In contrast to many existing learned MR image reconstruction techniques that necessitate coil-sensitivity map (CSM) estimation as a distinct network component, our proposed approach avoids explicit CSM estimation. Instead, it implicitly captures and learns to exploit the inter-coil relationships of the images. Our method consists of a series of novel learned image and k-space blocks with shared latent information and adaptation to the acquisition parameters by feature-wise modulation (FiLM), as well as coil-wise data-consistency (DC) blocks. Our method achieved PSNR values of 34.89 and 35.56 and SSIM values of 0.920 and 0.942 in the cine track and mapping track validation leaderboard of the MICCAI STACOM CMRxRecon Challenge, respectively, ranking 4th among different teams at the time of writing. Code will be made available at https://github.com/fzimmermann89/CMRxRecon

comment: Accepted at MICCAI STACOM 2023

♻ ☆ Navigating the Maze of Explainable AI: A Systematic Approach to Evaluating Methods and Metrics NeurIPS 2024

Lukas Klein, Carsten T. Lüth, Udo Schlegel, Till J. Bungert, Mennatallah El-Assady, Paul F. Jäger

Explainable AI (XAI) is a rapidly growing domain with a myriad of proposed methods as well as metrics aiming to evaluate their efficacy. However, current studies are often of limited scope, examining only a handful of XAI methods and ignoring underlying design parameters for performance, such as the model architecture or the nature of input data. Moreover, they often rely on one or a few metrics and neglect thorough validation, increasing the risk of selection bias and ignoring discrepancies among metrics. These shortcomings leave practitioners confused about which method to choose for their problem. In response, we introduce LATEC, a large-scale benchmark that critically evaluates 17 prominent XAI methods using 20 distinct metrics. We systematically incorporate vital design parameters like varied architectures and diverse input modalities, resulting in 7,560 examined combinations. Through LATEC, we showcase the high risk of conflicting metrics leading to unreliable rankings and consequently propose a more robust evaluation scheme. Further, we comprehensively evaluate various XAI methods to assist practitioners in selecting appropriate methods aligning with their needs. Curiously, the emerging top-performing method, Expected Gradients, is not examined in any relevant related study. LATEC reinforces its role in future XAI research by publicly releasing all 326k saliency maps and 378k metric scores as a (meta-)evaluation dataset. The benchmark is hosted at: https://github.com/IML-DKFZ/latec.

comment: Accepted at NeurIPS 2024

♻ ☆ Persistent Test-time Adaptation in Recurring Testing Scenarios NeurIPS 2024

Trung-Hieu Hoang, Duc Minh Vo, Minh N. Do

Current test-time adaptation (TTA) approaches aim to adapt to environments that change continuously. Yet, it is unclear whether TTA methods can maintain their adaptability over prolonged periods. To answer this question, we introduce a diagnostic setting - **recurring TTA** where environments not only change but also recur over time, creating an extensive data stream. This setting allows us to examine the error accumulation of TTA models, in the most basic scenario, when they are regularly exposed to previous testing environments. Furthermore, we simulate a TTA process on a simple yet representative $\epsilon$-**perturbed Gaussian Mixture Model Classifier**, deriving theoretical insights into the dataset- and algorithm-dependent factors contributing to gradual performance degradation. Our investigation leads us to propose **persistent TTA (PeTTA)**, which senses when the model is diverging towards collapse and adjusts the adaptation strategy, striking a balance between the dual objectives of adaptation and model collapse prevention. The supreme stability of PeTTA over existing approaches, in the face of lifelong TTA scenarios, has been demonstrated over comprehensive experiments on various benchmarks.

comment: Accepted to the 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

♻ ☆ Autoregressive Image Diffusion: Generation of Image Sequence and Application in MRI

Guanxiong Luo, Shoujin Huang, Martin Uecker

Magnetic resonance imaging (MRI) is a widely used non-invasive imaging modality. However, a persistent challenge lies in balancing image quality with imaging speed. This trade-off is primarily constrained by k-space measurements, which traverse specific trajectories in the spatial Fourier domain (k-space). These measurements are often undersampled to shorten acquisition times, resulting in image artifacts and compromised quality. Generative models learn image distributions and can be used to reconstruct high-quality images from undersampled k-space data. In this work, we present the autoregressive image diffusion (AID) model for image sequences and use it to sample the posterior for accelerated MRI reconstruction. The algorithm incorporates both undersampled k-space and pre-existing information. Models trained with fastMRI dataset are evaluated comprehensively. The results show that the AID model can robustly generate sequentially coherent image sequences. In MRI applications, the AID can outperform the standard diffusion model and reduce hallucinations, due to the learned inter-image dependencies. The project code is available at https://github.com/mrirecon/aid.

♻ ☆ WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models NeurIPS 2024

Peng Wang, Zexi Li, Ningyu Zhang, Ziwen Xu, Yunzhi Yao, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen

comment: NeurIPS 2024

♻ ☆ Selective Transformer for Hyperspectral Image Classification

Yichu Xu, Di Wang, Lefei Zhang, Liangpei Zhang

Transformer has achieved satisfactory results in the field of hyperspectral image (HSI) classification. However, existing Transformer models face two key challenges when dealing with HSI scenes characterized by diverse land cover types and rich spectral information: (1) fixed receptive field representation overlooks effective contextual information; (2) redundant self-attention feature representation. To address these limitations, we propose a novel Selective Transformer (SFormer) for HSI classification. The SFormer is designed to dynamically select receptive fields for capturing both spatial and spectral contextual information, while mitigating the impact of redundant data by prioritizing the most relevant features. This enables a highly accurate classification of the land covers of the HSI. Specifically, a Kernel Selective Transformer Block (KSTB) is first utilized to dynamically select an appropriate receptive field range to effectively extract spatial-spectral features. Furthermore, to capture the most crucial tokens, a Token Selective Transformer Block (TSTB) is introduced, which selects the most relevant tokens based on the ranking of attention scores for each query. Extensive experiments on four benchmark HSI datasets demonstrate that the proposed SFormer outperforms the state-of-the-art HSI classification models. The codes will be released.

♻ ☆ NeRAF: 3D Scene Infused Neural Radiance and Acoustic Fields

Amandine Brunetto, Sascha Hornauer, Fabien Moutarde

Sound plays a major role in human perception. Along with vision, it provides essential information for understanding our surroundings. Despite advances in neural implicit representations, learning acoustics that align with visual scenes remains a challenge. We propose NeRAF, a method that jointly learns acoustic and radiance fields. NeRAF synthesizes both novel views and spatialized room impulse responses (RIR) at new positions by conditioning the acoustic field on 3D scene geometric and appearance priors from the radiance field. The generated RIR can be applied to auralize any audio signal. Each modality can be rendered independently and at spatially distinct positions, offering greater versatility. We demonstrate that NeRAF generates high-quality audio on SoundSpaces and RAF datasets, achieving significant performance improvements over prior methods while being more data-efficient. Additionally, NeRAF enhances novel view synthesis of complex scenes trained with sparse data through cross-modal learning. NeRAF is designed as a Nerfstudio module, providing convenient access to realistic audio-visual generation.

comment: Project Page: https://amandinebtto.github.io/NeRAF

♻ ☆ IRASNet: Improved Feature-Level Clutter Reduction for Domain Generalized SAR-ATR

Oh-Tae Jang, Hae-Kang Song, Min-Jun Kim, Kyung-Hwan Lee, Geon Lee, Sung-Ho Kim, Hee-Sub Shin, Jae-Woo Ok, Min-Young Back, Jae-Hyuk Yoon, Kyung-Tae Kim

Recently, computer-aided design models and electromagnetic simulations have been used to augment synthetic aperture radar (SAR) data for deep learning. However, an automatic target recognition (ATR) model struggles with domain shift when using synthetic data because the model learns specific clutter patterns present in such data, which disturbs performance when applied to measured data with different clutter distributions. This study proposes a framework particularly designed for domain-generalized SAR-ATR called IRASNet, enabling effective feature-level clutter reduction and domain-invariant feature learning. First, we propose a clutter reduction module (CRM) that maximizes the signal-to-clutter ratio on feature maps. The module reduces the impact of clutter at the feature level while preserving target and shadow information, thereby improving ATR performance. Second, we integrate adversarial learning with CRM to extract clutter-reduced domain-invariant features. The integration bridges the gap between synthetic and measured datasets without requiring measured data during training. Third, we improve feature extraction from target and shadow regions by implementing a positional supervision task using mask ground truth encoding. The improvement enhances the ability of the model to discriminate between classes. Our proposed IRASNet presents new state-of-the-art public SAR datasets utilizing target and shadow information to achieve superior performance across various test conditions. IRASNet not only enhances generalization performance but also significantly improves feature-level clutter reduction, making it a valuable advancement in the field of radar image pattern recognition.

comment: 16 pages, 11 figures

♻ ☆ Boost Your NeRF: A Model-Agnostic Mixture of Experts Framework for High Quality and Efficient Rendering ECCV 2024

Francesco Di Sario, Riccardo Renzulli, Enzo Tartaglione, Marco Grangetto

Since the introduction of NeRFs, considerable attention has been focused on improving their training and inference times, leading to the development of Fast-NeRFs models. Despite demonstrating impressive rendering speed and quality, the rapid convergence of such models poses challenges for further improving reconstruction quality. Common strategies to improve rendering quality involves augmenting model parameters or increasing the number of sampled points. However, these computationally intensive approaches encounter limitations in achieving significant quality enhancements. This study introduces a model-agnostic framework inspired by Sparsely-Gated Mixture of Experts to enhance rendering quality without escalating computational complexity. Our approach enables specialization in rendering different scene components by employing a mixture of experts with varying resolutions. We present a novel gate formulation designed to maximize expert capabilities and propose a resolution-based routing technique to effectively induce sparsity and decompose scenes. Our work significantly improves reconstruction quality while maintaining competitive performance.

comment: The paper has been accepted to the ECCV 2024 conference

♻ ☆ Seeking Flat Minima with Mean Teacher on Semi- and Weakly-Supervised Domain Generalization for Object Detection

Ryosuke Furuta, Yoichi Sato

Object detectors do not work well when domains largely differ between training and testing data. To overcome this domain gap in object detection without requiring expensive annotations, we consider two problem settings: semi-supervised domain generalizable object detection (SS-DGOD) and weakly-supervised DGOD (WS-DGOD). In contrast to the conventional domain generalization for object detection that requires labeled data from multiple domains, SS-DGOD and WS-DGOD require labeled data only from one domain and unlabeled or weakly-labeled data from multiple domains for training. In this paper, we show that object detectors can be effectively trained on the two settings with the same Mean Teacher learning framework, where a student network is trained with pseudo-labels output from a teacher on the unlabeled or weakly-labeled data. We provide novel interpretations of why the Mean Teacher learning framework works well on the two settings in terms of the relationships between the generalization gap and flat minima in parameter space. On the basis of the interpretations, we also show that incorporating a simple regularization method into the Mean Teacher learning framework leads to flatter minima. The experimental results demonstrate that the regularization leads to flatter minima and boosts the performance of the detectors trained with the Mean Teacher learning framework on the two settings.

♻ ☆ Breaking the Frame: Visual Place Recognition by Overlap Prediction

Tong Wei, Philipp Lindenberger, Jiri Matas, Daniel Barath

Visual place recognition methods struggle with occlusions and partial visual overlaps. We propose a novel visual place recognition approach based on overlap prediction, called VOP, shifting from traditional reliance on global image similarities and local features to image overlap prediction. VOP proceeds co-visible image sections by obtaining patch-level embeddings using a Vision Transformer backbone and establishing patch-to-patch correspondences without requiring expensive feature detection and matching. Our approach uses a voting mechanism to assess overlap scores for potential database images. It provides a nuanced image retrieval metric in challenging scenarios. Experimental results show that VOP leads to more accurate relative pose estimation and localization results on the retrieved image pairs than state-of-the-art baselines on a number of large-scale, real-world indoor and outdoor benchmarks. The code is available at https://github.com/weitong8591/vop.git.

♻ ☆ Spectrum Extraction and Clipping for Implicitly Linear Layers

Ali Ebrahimpour Boroojeny, Matus Telgarsky, Hari Sundaram

We show the effectiveness of automatic differentiation in efficiently and correctly computing and controlling the spectrum of implicitly linear operators, a rich family of layer types including all standard convolutional and dense layers. We provide the first clipping method which is correct for general convolution layers, and illuminate the representational limitation that caused correctness issues in prior work. We study the effect of the batch normalization layers when concatenated with convolutional layers and show how our clipping method can be applied to their composition. By comparing the accuracy and performance of our algorithms to the state-of-the-art methods, using various experiments, we show they are more precise and efficient and lead to better generalization and adversarial robustness. We provide the code for using our methods at https://github.com/Ali-E/FastClip.

♻ ☆ GS-Hider: Hiding Messages into 3D Gaussian Splatting NeurIPS 2024

Xuanyu Zhang, Jiarui Meng, Runyi Li, Zhipei Xu, Yongbing Zhang, Jian Zhang

3D Gaussian Splatting (3DGS) has already become the emerging research focus in the fields of 3D scene reconstruction and novel view synthesis. Given that training a 3DGS requires a significant amount of time and computational cost, it is crucial to protect the copyright, integrity, and privacy of such 3D assets. Steganography, as a crucial technique for encrypted transmission and copyright protection, has been extensively studied. However, it still lacks profound exploration targeted at 3DGS. Unlike its predecessor NeRF, 3DGS possesses two distinct features: 1) explicit 3D representation; and 2) real-time rendering speeds. These characteristics result in the 3DGS point cloud files being public and transparent, with each Gaussian point having a clear physical significance. Therefore, ensuring the security and fidelity of the original 3D scene while embedding information into the 3DGS point cloud files is an extremely challenging task. To solve the above-mentioned issue, we first propose a steganography framework for 3DGS, dubbed GS-Hider, which can embed 3D scenes and images into original GS point clouds in an invisible manner and accurately extract the hidden messages. Specifically, we design a coupled secured feature attribute to replace the original 3DGS's spherical harmonics coefficients and then use a scene decoder and a message decoder to disentangle the original RGB scene and the hidden message. Extensive experiments demonstrated that the proposed GS-Hider can effectively conceal multimodal messages without compromising rendering quality and possesses exceptional security, robustness, capacity, and flexibility. Our project is available at: https://xuanyuzhang21.github.io/project/gshider.

comment: Accepted by NeurIPS 2024, 3DGS steganography

♻ ☆ SKT: Integrating State-Aware Keypoint Trajectories with Vision-Language Models for Robotic Garment Manipulation

Xin Li, Siyuan Huang, Qiaojun Yu, Zhengkai Jiang, Ce Hao, Yimeng Zhu, Hongsheng Li, Peng Gao, Cewu Lu

Automating garment manipulation poses a significant challenge for assistive robotics due to the diverse and deformable nature of garments. Traditional approaches typically require separate models for each garment type, which limits scalability and adaptability. In contrast, this paper presents a unified approach using vision-language models (VLMs) to improve keypoint prediction across various garment categories. By interpreting both visual and semantic information, our model enables robots to manage different garment states with a single model. We created a large-scale synthetic dataset using advanced simulation techniques, allowing scalable training without extensive real-world data. Experimental results indicate that the VLM-based method significantly enhances keypoint detection accuracy and task success rates, providing a more flexible and general solution for robotic garment manipulation. In addition, this research also underscores the potential of VLMs to unify various garment manipulation tasks within a single framework, paving the way for broader applications in home automation and assistive robotics for future.

♻ ☆ Visual Question Decomposition on Multimodal Large Language Models EMNLP2024

Haowei Zhang, Jianzhe Liu, Zhen Han, Shuo Chen, Bailan He, Volker Tresp, Zhiqiang Xu, Jindong Gu

comment: Accepted to EMNLP2024 Findings

♻ ☆ StructuReiser: A Structure-preserving Video Stylization Method

Radim Spetlik, David Futschik, Daniel Sykora

We introduce StructuReiser, a novel video-to-video translation method that transforms input videos into stylized sequences using a set of user-provided keyframes. Unlike existing approaches, StructuReiser maintains strict adherence to the structural elements of the target video, preserving the original identity while seamlessly applying the desired stylistic transformations. This enables a level of control and consistency that was previously unattainable with traditional text-driven or keyframe-based methods. Furthermore, StructuReiser supports real-time inference and custom keyframe editing, making it ideal for interactive applications and expanding the possibilities for creative expression and video manipulation.

♻ ☆ VILENS: Visual, Inertial, Lidar, and Leg Odometry for All-Terrain Legged Robots

David Wisth, Marco Camurri, Maurice Fallon

We present visual inertial lidar legged navigation system (VILENS), an odometry system for legged robots based on factor graphs. The key novelty is the tight fusion of four different sensor modalities to achieve reliable operation when the individual sensors would otherwise produce degenerate estimation. To minimize leg odometry drift, we extend the robot's state with a linear velocity bias term, which is estimated online. This bias is observable because of the tight fusion of this preintegrated velocity factor with vision, lidar, and inertial measurement unit (IMU) factors. Extensive experimental validation on different ANYmal quadruped robots is presented, for a total duration of 2 h and 1.8 km traveled. The experiments involved dynamic locomotion over loose rocks, slopes, and mud, which caused challenges such as slippage and terrain deformation. Perceptual challenges included dark and dusty underground caverns, and open and feature-deprived areas. We show an average improvement of 62% translational and 51% rotational errors compared to a state-of-the-art loosely coupled approach. To demonstrate its robustness, VILENS was also integrated with a perceptive controller and a local path planner.

comment: Video: https://youtu.be/NG4pkjJKhus

♻ ☆ Generalized Consistency Trajectory Models for Image Manipulation

Beomsu Kim, Jaemin Kim, Jeongsol Kim, Jong Chul Ye

Diffusion models (DMs) excel in unconditional generation, as well as on applications such as image editing and restoration. The success of DMs lies in the iterative nature of diffusion: diffusion breaks down the complex process of mapping noise to data into a sequence of simple denoising tasks. Moreover, we are able to exert fine-grained control over the generation process by injecting guidance terms into each denoising step. However, the iterative process is also computationally intensive, often taking from tens up to thousands of function evaluations. Although consistency trajectory models (CTMs) enable traversal between any time points along the probability flow ODE (PFODE) and score inference with a single function evaluation, CTMs only allow translation from Gaussian noise to data. This work aims to unlock the full potential of CTMs by proposing generalized CTMs (GCTMs), which translate between arbitrary distributions via ODEs. We discuss the design space of GCTMs and demonstrate their efficacy in various image manipulation tasks such as image-to-image translation, restoration, and editing.

♻ ☆ Learning Contrastive Feature Representations for Facial Action Unit Detection

Ziqiao Shang, Bin Liu, Fengmao Lv, Fei Teng, Tianrui Li

Facial action unit (AU) detection has long encountered the challenge of detecting subtle feature differences when AUs activate. Existing methods often rely on encoding pixel-level information of AUs, which not only encodes additional redundant information but also leads to increased model complexity and limited generalizability. Additionally, the accuracy of AU detection is negatively impacted by the class imbalance issue of each AU type, and the presence of noisy and false AU labels. In this paper, we introduce a novel contrastive learning framework aimed for AU detection that incorporates both self-supervised and supervised signals, thereby enhancing the learning of discriminative features for accurate AU detection. To tackle the class imbalance issue, we employ a negative sample re-weighting strategy that adjusts the step size of updating parameters for minority and majority class samples. Moreover, to address the challenges posed by noisy and false AU labels, we employ a sampling technique that encompasses three distinct types of positive sample pairs. This enables us to inject self-supervised signals into the supervised signal, effectively mitigating the adverse effects of noisy labels. Our experimental assessments, conducted on four widely-utilized benchmark datasets (BP4D, DISFA, GFT and Aff-Wild2), underscore the superior performance of our approach compared to state-of-the-art methods of AU detection. Our code is available at \url{https://github.com/Ziqiao-Shang/AUNCE}.

comment: 13 pages, 17 figures, submitted to IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

♻ ☆ Classification of All Blood Cell Images using ML and DL Models

Rabia Asghar, Sanjay Kumar, Paul Hynds, Abeera Mahfooz

Human blood primarily comprises plasma, red blood cells, white blood cells, and platelets. It plays a vital role in transporting nutrients to different organs, where it stores essential health-related data about the human body. Blood cells are utilized to defend the body against diverse infections, including fungi, viruses, and bacteria. Hence, blood analysis can help physicians assess an individual's physiological condition. Blood cells have been sub-classified into eight groups: Neutrophils, eosinophils, basophils, lymphocytes, monocytes, immature granulocytes (promyelocytes, myelocytes, and metamyelocytes), erythroblasts, and platelets or thrombocytes on the basis of their nucleus, shape, and cytoplasm. Traditionally, pathologists and hematologists in laboratories have examined these blood cells using a microscope before manually classifying them. The manual approach is slower and more prone to human error. Therefore, it is essential to automate this process. In our paper, transfer learning with CNN pre-trained models. VGG16, VGG19, ResNet-50, ResNet-101, ResNet-152, InceptionV3, MobileNetV2, and DenseNet-20 applied to the PBC dataset's normal DIB. The overall accuracy achieved with these models lies between 91.375 and 94.72%. Hence, inspired by these pre-trained architectures, a model has been proposed to automatically classify the ten types of blood cells with increased accuracy. A novel CNN-based framework has been presented to improve accuracy. The proposed CNN model has been tested on the PBC dataset normal DIB. The outcomes of the experiments demonstrate that our CNN-based framework designed for blood cell classification attains an accuracy of 99.91% on the PBC dataset. Our proposed convolutional neural network model performs competitively when compared to earlier results reported in the literature.

comment: 15

♻ ☆ A Framework for Pupil Tracking with Event Cameras

Khadija Iddrisu, Waseem Shariff, Suzanne Little

Saccades are extremely rapid movements of both eyes that occur simultaneously, typically observed when an individual shifts their focus from one object to another. These movements are among the swiftest produced by humans and possess the potential to achieve velocities greater than that of blinks. The peak angular speed of the eye during a saccade can reach as high as 700{\deg}/s in humans, especially during larger saccades that cover a visual angle of 25{\deg}. Previous research has demonstrated encouraging outcomes in comprehending neurological conditions through the study of saccades. A necessary step in saccade detection involves accurately identifying the precise location of the pupil within the eye, from which additional information such as gaze angles can be inferred. Conventional frame-based cameras often struggle with the high temporal precision necessary for tracking very fast movements, resulting in motion blur and latency issues. Event cameras, on the other hand, offer a promising alternative by recording changes in the visual scene asynchronously and providing high temporal resolution and low latency. By bridging the gap between traditional computer vision and event-based vision, we present events as frames that can be readily utilized by standard deep learning algorithms. This approach harnesses YOLOv8, a state-of-the-art object detection technology, to process these frames for pupil tracking using the publicly accessible Ev-Eye dataset. Experimental results demonstrate the framework's effectiveness, highlighting its potential applications in neuroscience, ophthalmology, and human-computer interaction.

comment: This paper is a preprint of a paper submitted to the 26th Irish Machine Vision and Image Processing Conference (IMVIP 2024). If accepted, the copy of record will be available at IET Digital Library

♻ ☆ Universal Medical Image Representation Learning with Compositional Decoders

Kaini Wang, Ling Yang, Siping Zhou, Guangquan Zhou, Wentao Zhang, Bin Cui, Shuo Li

Visual-language models have advanced the development of universal models, yet their application in medical imaging remains constrained by specific functional requirements and the limited data. Current general-purpose models are typically designed with task-specific branches and heads, which restricts the shared feature space and the flexibility of model. To address these challenges, we have developed a decomposed-composed universal medical imaging paradigm (UniMed) that supports tasks at all levels. To this end, we first propose a decomposed decoder that can predict two types of outputs -- pixel and semantic, based on a defined input queue. Additionally, we introduce a composed decoder that unifies the input and output spaces and standardizes task annotations across different levels into a discrete token format. The coupled design of these two components enables the model to flexibly combine tasks and mutual benefits. Moreover, our joint representation learning strategy skilfully leverages large amounts of unlabeled data and unsupervised loss, achieving efficient one-stage pretraining for more robust performance. Experimental results show that UniMed achieves state-of-the-art performance on eight datasets across all three tasks and exhibits strong zero-shot and 100-shot transferability. We will release the code and trained models upon the paper's acceptance.

♻ ☆ ColorwAI: Generative Colorways of Textiles through GAN and Diffusion Disentanglement ECCV 2024

Ludovica Schaerf, Andrea Alfarano, Eric Postma

Colorway creation is the task of generating textile samples in alternate color variations maintaining an underlying pattern. The individuation of a suitable color palette for a colorway is a complex creative task, responding to client and market needs, stylistic and cultural specifications, and mood. We introduce a modification of this task, the "generative colorway" creation, that includes minimal shape modifications, and propose a framework, "ColorwAI", to tackle this task using color disentanglement on StyleGAN and Diffusion. We introduce a variation of the InterfaceGAN method for supervised disentanglement, ShapleyVec. We use Shapley values to subselect a few dimensions of the detected latent direction. Moreover, we introduce a general framework to adopt common disentanglement methods on any architecture with a semantic latent space and test it on Diffusion and GANs. We interpret the color representations within the models' latent space. We find StyleGAN's W space to be the most aligned with human notions of color. Finally, we suggest that disentanglement can solicit a creative system for colorway creation, and evaluate it through expert questionnaires and creativity theory.

comment: Accepted to ECCV 2024 VISART workshop, oral presentation

♻ ☆ Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training NeurIPS 2024

Haoran He, Chenjia Bai, Ling Pan, Weinan Zhang, Bin Zhao, Xuelong Li

Learning a generalist embodied agent capable of completing multiple tasks poses challenges, primarily stemming from the scarcity of action-labeled robotic datasets. In contrast, a vast amount of human videos exist, capturing intricate tasks and interactions with the physical world. Promising prospects arise for utilizing actionless human videos for pre-training and transferring the knowledge to facilitate robot policy learning through limited robot demonstrations. However, it remains a challenge due to the domain gap between humans and robots. Moreover, it is difficult to extract useful information representing the dynamic world from human videos, because of its noisy and multimodal data structure. In this paper, we introduce a novel framework to tackle these challenges, which leverages a unified discrete diffusion to combine generative pre-training on human videos and policy fine-tuning on a small number of action-labeled robot videos. We start by compressing both human and robot videos into unified video tokens. In the pre-training stage, we employ a discrete diffusion model with a mask-and-replace diffusion strategy to predict future video tokens in the latent space. In the fine-tuning stage, we harness the imagined future videos to guide low-level action learning with a limited set of robot data. Experiments demonstrate that our method generates high-fidelity future videos for planning and enhances the fine-tuned policies compared to previous state-of-the-art approaches with superior performance. Our project website is available at https://video-diff.github.io/.

comment: Accepted by NeurIPS 2024. 24 pages

♻ ☆ Can Large Language Models Understand Symbolic Graphics Programs?

Zeju Qiu, Weiyang Liu, Haiwen Feng, Zhen Liu, Tim Z. Xiao, Katherine M. Collins, Joshua B. Tenenbaum, Adrian Weller, Michael J. Black, Bernhard Schölkopf

comment: Technical Report v2 (46 pages, 24 figures, project page: https://sgp-bench.github.io/, substantial update from v1)

♻ ☆ TD-NeRF: Novel Truncated Depth Prior for Joint Camera Pose and Neural Radiance Field Optimization

Zhen Tan, Zongtan Zhou, Yangbing Ge, Zi Wang, Xieyuanli Chen, Dewen Hu

The reliance on accurate camera poses is a significant barrier to the widespread deployment of Neural Radiance Fields (NeRF) models for 3D reconstruction and SLAM tasks. The existing method introduces monocular depth priors to jointly optimize the camera poses and NeRF, which fails to fully exploit the depth priors and neglects the impact of their inherent noise. In this paper, we propose Truncated Depth NeRF (TD-NeRF), a novel approach that enables training NeRF from unknown camera poses - by jointly optimizing learnable parameters of the radiance field and camera poses. Our approach explicitly utilizes monocular depth priors through three key advancements: 1) we propose a novel depth-based ray sampling strategy based on the truncated normal distribution, which improves the convergence speed and accuracy of pose estimation; 2) to circumvent local minima and refine depth geometry, we introduce a coarse-to-fine training strategy that progressively improves the depth precision; 3) we propose a more robust inter-frame point constraint that enhances robustness against depth noise during training. The experimental results on three datasets demonstrate that TD-NeRF achieves superior performance in the joint optimization of camera pose and NeRF, surpassing prior works, and generates more accurate depth geometry. The implementation of our method has been released at https://github.com/nubot-nudt/TD-NeRF.

♻ ☆ Exploring Distortion Prior with Latent Diffusion Models for Remote Sensing Image Compression

Junhui Li, Jutao Li, Xingsong Hou, Huake Wang

Deep learning-based image compression algorithms typically focus on designing encoding and decoding networks and improving the accuracy of entropy model estimation to enhance the rate-distortion (RD) performance. However, few algorithms leverage the compression distortion prior from existing compression algorithms to improve RD performance. In this paper, we propose a latent diffusion model-based remote sensing image compression (LDM-RSIC) method, which aims to enhance the final decoding quality of RS images by utilizing the generated distortion prior from a LDM. Our approach consists of two stages. In the first stage, a self-encoder learns prior from the high-quality input image. In the second stage, the prior is generated through an LDM, conditioned on the decoded image of an existing learning-based image compression algorithm, to be used as auxiliary information for generating the texture-rich enhanced image. To better utilize the prior, a channel attention and gate-based dynamic feature attention module (DFAM) is embedded into a Transformer-based multi-scale enhancement network (MEN) for image enhancement. Extensive experiments demonstrate the proposed LDM-RSIC significantly outperforms existing state-of-the-art traditional and learning-based image compression algorithms in terms of both subjective perception and objective metrics. Additionally, we use the LDM-based scheme to improve the traditional image compression algorithm JPEG2000 and obtain 32.00% bit savings on the DOTA testing set. The code will be available at https://github.com/mlkk518/LDM-RSIC.

♻ ☆ Greit-HRNet: Grouped Lightweight High-Resolution Network for Human Pose Estimation

Junjia Han

As multi-scale features are necessary for human pose estimation tasks, high-resolution networks are widely applied. To improve efficiency, lightweight modules are proposed to replace costly point-wise convolutions in high-resolution networks, including channel weighting and spatial weighting methods. However, they fail to maintain the consistency of weights and capture global spatial information. To address these problems, we present a Grouped lightweight High-Resolution Network (Greit-HRNet), in which we propose a Greit block including a group method Grouped Channel Weighting (GCW) and a spatial weighting method Global Spatial Weighting (GSW). GCW modules group conditional channel weighting to make weights stable and maintain the high-resolution features with the deepening of the network, while GSW modules effectively extract global spatial information and exchange information across channels. In addition, we apply the Large Kernel Attention (LKA) method to improve the whole efficiency of our Greit-HRNet. Our experiments on both MS-COCO and MPII human pose estimation datasets demonstrate the superior performance of our Greit-HRNet, outperforming other state-of-the-art lightweight networks.

comment: 16 pages, 4 figures

♻ ☆ MedThink: Explaining Medical Visual Question Answering via Multimodal Decision-Making Rationale

Xiaotang Gai, Chenyi Zhou, Jiaxiang Liu, Yang Feng, Jian Wu, Zuozhu Liu

Medical Visual Question Answering (MedVQA), which offers language responses to image-based medical inquiries, represents a challenging task and significant advancement in healthcare. It assists medical experts to swiftly interpret medical images, thereby enabling faster and more accurate diagnoses. However, the model interpretability and transparency of existing MedVQA solutions are often limited, posing challenges in understanding their decision-making processes. To address this issue, we devise a semi-automated annotation process to streamline data preparation and build new benchmark MedVQA datasets R-RAD, R-SLAKE and R-Path. These datasets provide intermediate medical decision-making rationales generated by multimodal large language models and human annotations for question-answering pairs in existing MedVQA datasets, i.e., VQA-RAD, SLAKE and PathVQA. Moreover, we design a novel framework, MedThink, which finetunes lightweight pretrained generative models by incorporating medical decision-making rationales. MedThink includes three distinct strategies to generate decision outcomes and corresponding rationales, thereby clearly showcasing the medical decision-making process during reasoning. Our comprehensive experiments show that our method achieves an accuracy of 83.5% on R-RAD, 86.3% on R-SLAKE and 87.2% on R-Path. These results significantly exceed those of existing state-of-the-art models with comparable parameters. Datasets and code will be released.

♻ ☆ 1st Place Solution to the 8th HANDS Workshop Challenge -- ARCTIC Track: 3DGS-based Bimanual Category-agnostic Interaction Reconstruction

Jeongwan On, Kyeonghwan Gwak, Gunyoung Kang, Hyein Hwang, Soohyun Hwang, Junuk Cha, Jaewook Han, Seungryul Baek

This report describes our 1st place solution to the 8th HANDS workshop challenge (ARCTIC track) in conjunction with ECCV 2024. In this challenge, we address the task of bimanual category-agnostic hand-object interaction reconstruction, which aims to generate 3D reconstructions of both hands and the object from a monocular video, without relying on predefined templates. This task is particularly challenging due to the significant occlusion and dynamic contact between the hands and the object during bimanual manipulation. We worked to resolve these issues by introducing a mask loss and a 3D contact loss, respectively. Moreover, we applied 3D Gaussian Splatting (3DGS) to this task. As a result, our method achieved a value of 38.69 in the main metric, CD$_h$, on the ARCTIC test set.

♻ ☆ ColPali: Efficient Document Retrieval with Vision Language Models

Manuel Faysse, Hugues Sibille, Tony Wu, Bilel Omrani, Gautier Viaud, Céline Hudelot, Pierre Colombo

comment: Under Review

♻ ☆ SARatrX: Towards Building A Foundation Model for SAR Target Recognition

Weijie Li, Wei Yang, Yuenan Hou, Li Liu, Yongxiang Liu, Xiang Li

Despite the remarkable progress in synthetic aperture radar automatic target recognition (SAR ATR), recent efforts have concentrated on the detection or classification of a specific and coarse category, e.g., vehicles, ships, airplanes, or buildings. One of the fundamental limitations of the top-performing SAR ATR methods is that the learning paradigm is supervised, task-specific, limited-category, closed-world learning, which depends on massive amounts of accurately annotated samples that are expensively labeled by expert SAR analysts and has limited generalization capability and scalability. In this work, we make the first attempt towards building a foundation model for SAR ATR, termed SARatrX. SARatrX learns generalizable representations via self-supervised learning (SSL) and provides a basis for label-efficient model adaptation to generic SAR target detection and classification tasks. Specifically, SARatrX is trained on 0.18 M unlabelled SAR target samples, which are curated by combining contemporary benchmarks and constitute the largest publicly available dataset till now. Considering the characteristics of SAR images, a backbone tailored for SAR ATR is carefully designed, and a two-step SSL method endowed with multi-scale gradient features was applied to ensure the feature diversity and model scalability of SARatrX. The capabilities of SARatrX are evaluated on classification under few-shot and robustness settings and detection across various categories and scenes, and impressive performance is achieved, often competitive with or even superior to prior fully supervised, semi-supervised, or self-supervised algorithms. Our SARatrX and the curated dataset are released at https://github.com/waterdisappear/SARatrX to foster research into foundation models for SAR ATR and SAR image interpretation.

♻ ☆ Expanding Scene Graph Boundaries: Fully Open-vocabulary Scene Graph Generation via Visual-Concept Alignment and Retention

Zuyao Chen, Jinlin Wu, Zhen Lei, Zhaoxiang Zhang, Changwen Chen

Scene Graph Generation (SGG) offers a structured representation critical in many computer vision applications. Traditional SGG approaches, however, are limited by a closed-set assumption, restricting their ability to recognize only predefined object and relation categories. To overcome this, we categorize SGG scenarios into four distinct settings based on the node and edge: Closed-set SGG, Open Vocabulary (object) Detection-based SGG (OvD-SGG), Open Vocabulary Relation-based SGG (OvR-SGG), and Open Vocabulary Detection + Relationbased SGG (OvD+R-SGG). While object-centric open vocabulary SGG has been studied recently, the more challenging problem of relation-involved open-vocabulary SGG remains relatively unexplored. To fill this gap, we propose a unified framework named OvSGTR towards fully open vocabulary SGG from a holistic view. The proposed framework is an end-to-end transformer architecture, which learns a visual-concept alignment for both nodes and edges, enabling the model to recognize unseen categories. For the more challenging settings of relation-involved open vocabulary SGG, the proposed approach integrates relation-aware pretraining utilizing image-caption data and retains visual-concept alignment through knowledge distillation. Comprehensive experimental results on the Visual Genome benchmark demonstrate the effectiveness and superiority of the proposed framework. Our code is available at https://github.com/gpt4vision/OvSGTR/.

♻ ☆ Discrete Distribution Networks

Lei Yang

We introduce a novel generative model, the Discrete Distribution Networks (DDN), that approximates data distribution using hierarchical discrete distributions. We posit that since the features within a network inherently capture distributional information, enabling the network to generate multiple samples simultaneously, rather than a single output, may offer an effective way to represent distributions. Therefore, DDN fits the target distribution, including continuous ones, by generating multiple discrete sample points. To capture finer details of the target data, DDN selects the output that is closest to the Ground Truth (GT) from the coarse results generated in the first layer. This selected output is then fed back into the network as a condition for the second layer, thereby generating new outputs more similar to the GT. As the number of DDN layers increases, the representational space of the outputs expands exponentially, and the generated samples become increasingly similar to the GT. This hierarchical output pattern of discrete distributions endows DDN with unique property: more general zero-shot conditional generation. We demonstrate the efficacy of DDN and its intriguing properties through experiments on CIFAR-10 and FFHQ. The code is available at https://discrete-distribution-networks.github.io/

comment: TL;DR: A Novel Generative Model with Simple Principles and Unique Properties

♻ ☆ Lighthouse: A User-Friendly Library for Reproducible Video Moment Retrieval and Highlight Detection EMNLP2024

Taichi Nishimura, Shota Nakada, Hokuto Munakata, Tatsuya Komatsu

comment: accepted at EMNLP2024 - system demonstration track

♻ ☆ PhoCoLens: Photorealistic and Consistent Reconstruction in Lensless Imaging NeurIPS 2024

Xin Cai, Zhiyuan You, Hailong Zhang, Wentao Liu, Jinwei Gu, Tianfan Xue

Lensless cameras offer significant advantages in size, weight, and cost compared to traditional lens-based systems. Without a focusing lens, lensless cameras rely on computational algorithms to recover the scenes from multiplexed measurements. However, current algorithms struggle with inaccurate forward imaging models and insufficient priors to reconstruct high-quality images. To overcome these limitations, we introduce a novel two-stage approach for consistent and photorealistic lensless image reconstruction. The first stage of our approach ensures data consistency by focusing on accurately reconstructing the low-frequency content with a spatially varying deconvolution method that adjusts to changes in the Point Spread Function (PSF) across the camera's field of view. The second stage enhances photorealism by incorporating a generative prior from pre-trained diffusion models. By conditioning on the low-frequency content retrieved in the first stage, the diffusion model effectively reconstructs the high-frequency details that are typically lost in the lensless imaging process, while also maintaining image fidelity. Our method achieves a superior balance between data fidelity and visual quality compared to existing methods, as demonstrated with two popular lensless systems, PhlatCam and DiffuserCam. Project website: https://phocolens.github.io/.

comment: NeurIPS 2024 Spotlight

♻ ☆ PhysDreamer: Physics-Based Interaction with 3D Objects via Video Generation ECCV 2024

Tianyuan Zhang, Hong-Xing Yu, Rundi Wu, Brandon Y. Feng, Changxi Zheng, Noah Snavely, Jiajun Wu, William T. Freeman

Realistic object interactions are crucial for creating immersive virtual experiences, yet synthesizing realistic 3D object dynamics in response to novel interactions remains a significant challenge. Unlike unconditional or text-conditioned dynamics generation, action-conditioned dynamics requires perceiving the physical material properties of objects and grounding the 3D motion prediction on these properties, such as object stiffness. However, estimating physical material properties is an open problem due to the lack of material ground-truth data, as measuring these properties for real objects is highly difficult. We present PhysDreamer, a physics-based approach that endows static 3D objects with interactive dynamics by leveraging the object dynamics priors learned by video generation models. By distilling these priors, PhysDreamer enables the synthesis of realistic object responses to novel interactions, such as external forces or agent manipulations. We demonstrate our approach on diverse examples of elastic objects and evaluate the realism of the synthesized interactions through a user study. PhysDreamer takes a step towards more engaging and realistic virtual experiences by enabling static 3D objects to dynamically respond to interactive stimuli in a physically plausible manner. See our project page at https://physdreamer.github.io/.

comment: Project website at: https://physdreamer.github.io/ Appear on ECCV 2024

♻ ☆ Deep Height Decoupling for Precise Vision-based 3D Occupancy Prediction

Yuan Wu, Zhiqiang Yan, Zhengxue Wang, Xiang Li, Le Hui, Jian Yang

The task of vision-based 3D occupancy prediction aims to reconstruct 3D geometry and estimate its semantic classes from 2D color images, where the 2D-to-3D view transformation is an indispensable step. Most previous methods conduct forward projection, such as BEVPooling and VoxelPooling, both of which map the 2D image features into 3D grids. However, the current grid representing features within a certain height range usually introduces many confusing features that belong to other height ranges. To address this challenge, we present Deep Height Decoupling (DHD), a novel framework that incorporates explicit height prior to filter out the confusing features. Specifically, DHD first predicts height maps via explicit supervision. Based on the height distribution statistics, DHD designs Mask Guided Height Sampling (MGHS) to adaptively decouple the height map into multiple binary masks. MGHS projects the 2D image features into multiple subspaces, where each grid contains features within reasonable height ranges. Finally, a Synergistic Feature Aggregation (SFA) module is deployed to enhance the feature representation through channel and spatial affinities, enabling further occupancy refinement. On the popular Occ3D-nuScenes benchmark, our method achieves state-of-the-art performance even with minimal input frames. Code is available at https://github.com/yanzq95/DHD.

♻ ☆ Correcting Diffusion Generation through Resampling

Yujian Liu, Yang Zhang, Tommi Jaakkola, Shiyu Chang

Despite diffusion models' superior capabilities in modeling complex distributions, there are still non-trivial distributional discrepancies between generated and ground-truth images, which has resulted in several notable problems in image generation, including missing object errors in text-to-image generation and low image quality. Existing methods that attempt to address these problems mostly do not tend to address the fundamental cause behind these problems, which is the distributional discrepancies, and hence achieve sub-optimal results. In this paper, we propose a particle filtering framework that can effectively address both problems by explicitly reducing the distributional discrepancies. Specifically, our method relies on a set of external guidance, including a small set of real images and a pre-trained object detector, to gauge the distribution gap, and then design the resampling weight accordingly to correct the gap. Experiments show that our methods can effectively correct missing object errors and improve image quality in various image generation tasks. Notably, our method outperforms the existing strongest baseline by 5% in object occurrence and 1.0 in FID on MS-COCO. Our code is publicly available at https://github.com/UCSB-NLP-Chang/diffusion_resampling.git.

♻ ☆ SplatSim: Zero-Shot Sim2Real Transfer of RGB Manipulation Policies Using Gaussian Splatting

Mohammad Nomaan Qureshi, Sparsh Garg, Francisco Yandun, David Held, George Kantor, Abhisesh Silwal

Sim2Real transfer, particularly for manipulation policies relying on RGB images, remains a critical challenge in robotics due to the significant domain shift between synthetic and real-world visual data. In this paper, we propose SplatSim, a novel framework that leverages Gaussian Splatting as the primary rendering primitive to reduce the Sim2Real gap for RGB-based manipulation policies. By replacing traditional mesh representations with Gaussian Splats in simulators, SplatSim produces highly photorealistic synthetic data while maintaining the scalability and cost-efficiency of simulation. We demonstrate the effectiveness of our framework by training manipulation policies within SplatSim and deploying them in the real world in a zero-shot manner, achieving an average success rate of 86.25%, compared to 97.5% for policies trained on real-world data. Videos can be found on our project page: https://splatsim.github.io

♻ ☆ iSeg: An Iterative Refinement-based Framework for Training-free Segmentation

Lin Sun, Jiale Cao, Jin Xie, Fahad Shahbaz Khan, Yanwei Pang

comment: Project Page: https://linsun449.github.io/iSeg/ Code: https://github.com/linsun449/iseg.code

♻ ☆ Perceptual Piercing: Human Visual Cue-based Object Detection in Low Visibility Conditions

Ashutosh Kumar

This study proposes a novel deep learning framework inspired by atmospheric scattering and human visual cortex mechanisms to enhance object detection under poor visibility scenarios such as fog, smoke, and haze. These conditions pose significant challenges for object recognition, impacting various sectors, including autonomous driving, aviation management, and security systems. The objective is to enhance the precision and reliability of detection systems under adverse environmental conditions. The research investigates the integration of human-like visual cues, particularly focusing on selective attention and environmental adaptability, to ascertain their impact on object detection's computational efficiency and accuracy. This paper proposes a multi-tiered strategy that integrates an initial quick detection process, followed by targeted region-specific dehazing, and concludes with an in-depth detection phase. The approach is validated using the Foggy Cityscapes, RESIDE-beta (OTS and RTTS) datasets and is anticipated to set new performance standards in detection accuracy while significantly optimizing computational efficiency. The findings offer a viable solution for enhancing object detection in poor visibility and contribute to the broader understanding of integrating human visual principles into deep learning algorithms for intricate visual recognition challenges.

♻ ☆ A Survey on Trustworthiness in Foundation Models for Medical Image Analysis

Congzhen Shi, Ryan Rezai, Jiaxi Yang, Qi Dou, Xiaoxiao Li

The rapid advancement of foundation models in medical imaging represents a significant leap toward enhancing diagnostic accuracy and personalized treatment. However, the deployment of foundation models in healthcare necessitates a rigorous examination of their trustworthiness, encompassing privacy, robustness, reliability, explainability, and fairness. The current body of survey literature on foundation models in medical imaging reveals considerable gaps, particularly in the area of trustworthiness. Additionally, existing surveys on the trustworthiness of foundation models do not adequately address their specific variations and applications within the medical imaging domain. This survey aims to fill that gap by presenting a novel taxonomy of foundation models used in medical imaging and analyzing the key motivations for ensuring their trustworthiness. We review current research on foundation models in major medical imaging applications, focusing on segmentation, medical report generation, medical question and answering (Q\&A), and disease diagnosis. These areas are highlighted because they have seen a relatively mature and substantial number of foundation models compared to other applications. We focus on literature that discusses trustworthiness in medical image analysis manuscripts. We explore the complex challenges of building trustworthy foundation models for each application, summarizing current concerns and strategies for enhancing trustworthiness. Furthermore, we examine the potential of these models to revolutionize patient care. Our analysis underscores the imperative for advancing towards trustworthy AI in medical image analysis, advocating for a balanced approach that fosters innovation while ensuring ethical and equitable healthcare delivery.

♻ ☆ BDetCLIP: Multimodal Prompting Contrastive Test-Time Backdoor Detection

Yuwei Niu, Shuo He, Qi Wei, Zongyu Wu, Feng Liu, Lei Feng

Multimodal contrastive learning methods (e.g., CLIP) have shown impressive zero-shot classification performance due to their strong ability to joint representation learning for visual and textual modalities. However, recent research revealed that multimodal contrastive learning on poisoned pre-training data with a small proportion of maliciously backdoored data can induce backdoored CLIP that could be attacked by inserted triggers in downstream tasks with a high success rate. To defend against backdoor attacks on CLIP, existing defense methods focus on either the pre-training stage or the fine-tuning stage, which would unfortunately cause high computational costs due to numerous parameter updates. In this paper, we provide the first attempt at a computationally efficient backdoor detection method to defend against backdoored CLIP in the inference stage. We empirically find that the visual representations of backdoored images are insensitive to both benign and malignant changes in class description texts. Motivated by this observation, we propose BDetCLIP, a novel test-time backdoor detection method based on contrastive prompting. Specifically, we first prompt the language model (e.g., GPT-4) to produce class-related description texts (benign) and class-perturbed random texts (malignant) by specially designed instructions. Then, the distribution difference in cosine similarity between images and the two types of class description texts can be used as the criterion to detect backdoor samples. Extensive experiments validate that our proposed BDetCLIP is superior to state-of-the-art backdoor detection methods, in terms of both effectiveness and efficiency.

♻ ☆ InFusionSurf: Refining Neural RGB-D Surface Reconstruction Using Per-Frame Intrinsic Refinement and TSDF Fusion Prior Learning ICME'24

Seunghwan Lee, Gwanmo Park, Hyewon Son, Jiwon Ryu, Han Joo Chae

We introduce InFusionSurf, an innovative enhancement for neural radiance field (NeRF) frameworks in 3D surface reconstruction using RGB-D video frames. Building upon previous methods that have employed feature encoding to improve optimization speed, we further improve the reconstruction quality with minimal impact on optimization time by refining depth information. InFusionSurf addresses camera motion-induced blurs in each depth frame through a per-frame intrinsic refinement scheme. It incorporates the truncated signed distance field (TSDF) Fusion, a classical real-time 3D surface reconstruction method, as a pretraining tool for the feature grid, enhancing reconstruction details and training speed. Comparative quantitative and qualitative analyses show that InFusionSurf reconstructs scenes with high accuracy while maintaining optimization efficiency. The effectiveness of our intrinsic refinement and TSDF Fusion-based pretraining is further validated through an ablation study.

comment: ICME'24 (Oral), Project page: https://rokit-healthcare.github.io/InFusionSurf/

♻ ☆ SpinQuant: LLM quantization with learned rotations

Zechun Liu, Changsheng Zhao, Igor Fedorov, Bilge Soran, Dhruv Choudhary, Raghuraman Krishnamoorthi, Vikas Chandra, Yuandong Tian, Tijmen Blankevoort

♻ ☆ PACE: marrying generalization in PArameter-efficient fine-tuning with Consistency rEgularization NeurIPS 2024

Yao Ni, Shan Zhang, Piotr Koniusz

Parameter-Efficient Fine-Tuning (PEFT) effectively adapts pre-trained vision transformers to downstream tasks. However, the optimization for tasks performance often comes at the cost of generalizability in fine-tuned models. To address this issue, we theoretically connect smaller weight gradient norms during training and larger datasets to the improved model generalization. Motivated by this connection, we propose reducing gradient norms for enhanced generalization and aligning fine-tuned model with the pre-trained counterpart to retain knowledge from large-scale pre-training data. Yet, naive alignment does not guarantee gradient reduction and can potentially cause gradient explosion, complicating efforts to manage gradients. To address such issues, we propose PACE, marrying generalization of PArameter-efficient fine-tuning with Consistency rEgularization. We perturb features learned from the adapter with the multiplicative noise and ensure the fine-tuned model remains consistent for same sample under different perturbations. Theoretical analysis shows that PACE not only implicitly regularizes gradients for enhanced generalization, but also implicitly aligns the fine-tuned and pre-trained models to retain knowledge. Experimental evidence supports our theories. PACE outperforms existing PEFT methods in four visual adaptation tasks: VTAB-1k, FGVC, few-shot learning and domain adaptation. Code will be available at https://github.com/MaxwellYaoNi/PACE

comment: Accepted by NeurIPS 2024 as a spotlight. This preliminary version will soon be extended with the experiments and analyses from the rebuttal

Hardware Architecture 5

☆ Secure Software/Hardware Hybrid In-Field Testing for System-on-Chip

Saleh Mulhem, Christian Ewert, Andrija Neskovic, Amrit Sharma Poudel, Christoph Hübner, Mladen Berekovic, Rainer Buchty

Modern Systems-on-Chip (SoCs) incorporate built-in self-test (BIST) modules deeply integrated into the device's intellectual property (IP) blocks. Such modules handle hardware faults and defects during device operation. As such, BIST results potentially reveal the internal structure and state of the device under test (DUT) and hence open attack vectors. So-called result compaction can overcome this vulnerability by hiding the BIST chain structure but introduces the issues of aliasing and invalid signatures. Software-BIST provides a flexible solution, that can tackle these issues, but suffers from limited observability and fault coverage. In this paper, we hence introduce a low-overhead software/hardware hybrid approach that overcomes the mentioned limitations. It relies on (a) keyed-hash message authentication code (KMAC) available on the SoC providing device-specific secure and valid signatures with zero aliasing and (b) the SoC processor for test scheduling hence increasing DUT availability. The proposed approach offers both on-chip- and remote-testing capabilities. We showcase a RISC-V-based SoC to demonstrate our approach, discussing system overhead and resulting compaction rates.

comment: PREPRINT - To appear at IFIP/IEEE VLSI-SoC 2024

☆ HF-NTT: Hazard-Free Dataflow Accelerator for Number Theoretic Transform

Xiangchen Meng, Zijun Jiang, Yangdi Lyu

Polynomial multiplication is one of the fundamental operations in many applications, such as fully homomorphic encryption (FHE). However, the computational inefficiency stemming from polynomials with many large-bit coefficients poses a significant challenge for the practical implementation of FHE. The Number Theoretic Transform (NTT) has proven an effective tool in enhancing polynomial multiplication, but a fast and adaptable method for generating NTT accelerators is lacking. In this paper, we introduce HF-NTT, a novel NTT accelerator. HF-NTT efficiently handles polynomials of varying degrees and moduli, allowing for a balance between performance and hardware resources by adjusting the number of Processing Elements (PEs). Meanwhile, we introduce a data movement strategy that eliminates the need for bit-reversal operations, resolves different hazards, and reduces the clock cycles. Furthermore, Our accelerator includes a hardware-friendly modular multiplication design and a configurable PE capable of adapting its data path, resulting in a universal architecture. We synthesized and implemented prototype using Vivado 2022.2, and evaluated it on the Xilinx Virtex-7 FPGA platform. The results demonstrate significant improvements in Area-Time-Product (ATP) and processing speed for different polynomial degrees. In scenarios involving multi-modulus polynomial multiplication, our prototype consistently outperforms other designs in both ATP and latency metrics.

♻ ☆ LoopTree: Exploring the Fused-layer Dataflow Accelerator Design Space

Michael Gilbert, Yannan Nellie Wu, Joel S. Emer, Vivienne Sze

comment: To be published in IEEE Transactions on Circuits and Systems for Artificial Intelligence

♻ ☆ RTLCoder: Fully Open-Source and Efficient LLM-Assisted RTL Code Generation Technique

Shang Liu, Wenji Fang, Yao Lu, Jing Wang, Qijun Zhang, Hongce Zhang, Zhiyao Xie

comment: Accepted by IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

♻ ☆ IR-Aware ECO Timing Optimization Using Reinforcement Learning

Wenjing Jiang, Vidya A. Chhabria, Sachin S. Sapatnekar

Engineering change orders (ECOs) in late stages make minimal design fixes to recover from timing shifts due to excessive IR drops. This paper integrates IR-drop-aware timing analysis and ECO timing optimization using reinforcement learning (RL). The method operates after physical design and power grid synthesis, and rectifies IR-drop-induced timing degradation through gate sizing. It incorporates the Lagrangian relaxation (LR) technique into a novel RL framework, which trains a relational graph convolutional network (R-GCN) agent to sequentially size gates to fix timing violations. The R-GCN agent outperforms a classical LR-only algorithm: in an open 45nm technology, it (a) moves the Pareto front of the delay-power tradeoff curve to the left (b) saves runtime over the prior approaches by running fast inference using trained models, and (c) reduces the perturbation to placement by sizing fewer cells. The RL model is transferable across timing specifications and to unseen designs with fine tuning.

Distributed, Parallel, and Cluster Computing 11

☆ A Seesaw Model Attack Algorithm for Distributed Learning

Kun Yang, Tianyi Luo, Yanjie Dong, Aohan Li

We investigate the Byzantine attack problem within the context of model training in distributed learning systems. While ensuring the convergence of current model training processes, common solvers (e.g. SGD, Adam, RMSProp, etc.) can be easily compromised by malicious nodes in these systems. Consequently, the training process may either converge slowly or even diverge. To develop effective secure distributed learning solvers, it is crucial to first examine attack methods to assess the robustness of these solvers. In this work, we contribute to the design of attack strategies by initially highlighting the limitations of finite-norm attacks. We then introduce the seesaw attack, which has been demonstrated to be more effective than the finite-norm attack. Through numerical experiments, we evaluate the efficacy of the seesaw attack across various gradient aggregation rules.

comment: Accepted for presentation at IEEE SmartIoT 2024

☆ A Digital Twin Framework for Liquid-cooled Supercomputers as Demonstrated at Exascale

Wesley Brewer, Matthias Maiterth, Vineet Kumar, Rafal Wojda, Sedrick Bouknight, Jesse Hines, Woong Shin, Scott Greenwood, David Grant, Wesley Williams, Feiyi Wang

We present ExaDigiT, an open-source framework for developing comprehensive digital twins of liquid-cooled supercomputers. It integrates three main modules: (1) a resource allocator and power simulator, (2) a transient thermo-fluidic cooling model, and (3) an augmented reality model of the supercomputer and central energy plant. The framework enables the study of "what-if" scenarios, system optimizations, and virtual prototyping of future systems. Using Frontier as a case study, we demonstrate the framework's capabilities by replaying six months of system telemetry for systematic verification and validation. Such a comprehensive analysis of a liquid-cooled exascale supercomputer is the first of its kind. ExaDigiT elucidates complex transient cooling system dynamics, runs synthetic or real workloads, and predicts energy losses due to rectification and voltage conversion. Throughout our paper, we present lessons learned to benefit HPC practitioners developing similar digital twins. We envision the digital twin will be a key enabler for sustainable, energy-efficient supercomputing.

comment: 14 pages, 9 figures, To be published in the Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis. 2024

☆ DIMS: Distributed Index for Similarity Search in Metric Spaces

Yifan Zhu, Chengyang Luo, Tang Qian, Lu Chen, Yunjun Gao, Baihua Zheng

Similarity search finds objects that are similar to a given query object based on a similarity metric. As the amount and variety of data continue to grow, similarity search in metric spaces has gained significant attention. Metric spaces can accommodate any type of data and support flexible distance metrics, making similarity search in metric spaces beneficial for many real-world applications, such as multimedia retrieval, personalized recommendation, trajectory analytics, data mining, decision planning, and distributed servers. However, existing studies mostly focus on indexing metric spaces on a single machine, which faces efficiency and scalability limitations with increasing data volume and query amount. Recent advancements in similarity search turn towards distributed methods, while they face challenges including inefficient local data management, unbalanced workload, and low concurrent search efficiency. To this end, we propose DIMS, an efficient Distributed Index for similarity search in Metric Spaces. First, we design a novel three-stage heterogeneous partition to achieve workload balance. Then, we present an effective three-stage indexing structure to efficiently manage objects. We also develop concurrent search methods with filtering and validation techniques that support efficient distributed similarity search. Additionally, we devise a cost-based optimization model to balance communication and computation cost. Extensive experiments demonstrate that DIMS significantly outperforms existing distributed similarity search approaches.

☆ Fast State Restoration in LLM Serving with HCache EuroSys 2025

Shiwei Gao, Youmin Chen, Jiwu Shu

The growing complexity of LLM usage today, e.g., multi-round conversation and retrieval-augmented generation (RAG), makes contextual states (i.e., KV cache) reusable across user requests. Given the capacity constraints of GPU memory, only a limited number of contexts can be cached on GPU for reusing. Existing inference systems typically evict part of the KV cache and restore it by recomputing it from the original tokens or offloading it to host storage for later retrieval, both of which introduce substantial computational or I/O overheads. We propose HCache, a novel LLM state restoration method. Its key idea is to restore LLM states from intermediate activations and thus utilize computational and I/O resources with low overhead. We enhance HCache with two techniques, including i) a bubble-free restoration scheduler that integrates resource-complementary methods to optimize the balance between computation and IO tasks; and ii) a chunk-based storage manager to address the layout mismatch issue (i.e., layer-before-token saving versus token-before-layer restoration). Our evaluations, conducted using real-world tasks, show that HCache reduces the TTFT by up to 1.93X compared to KV offload while consuming 1.92-2.40X less storage space; compared to token recomputation, HCache achieves up to 5.73X reduction in TTFT.

comment: EuroSys 2025

☆ Cloud-Based Scheduling Mechanism for Scalable and Resource-Efficient Centralized Controllers

Achilleas Santi Seisa, Sumeet Gajanan Satpute, George Nikolakopoulos

This paper proposes a novel approach to address the challenges of deploying complex robotic software in large-scale systems, i.e., Centralized Nonlinear Model Predictive Controllers (CNMPCs) for multi-agent systems. The proposed approach is based on a Kubernetes-based scheduling mechanism designed to monitor and optimize the operation of CNMPCs, while addressing the scalability limitation of centralized control schemes. By leveraging a cluster in a real-time cloud environment, the proposed mechanism effectively offloads the computational burden of CNMPCs. Through experiments, we have demonstrated the effectiveness and performance of our system, especially in scenarios where the number of robots is subject to change. Our work contributes to the advancement of cloud-based control strategies and lays the foundation for enhanced performance in cloud-controlled robotic systems.

comment: 7 pages, 6 figures, IECON 2024

☆ FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

Haokun Chen, Hang Li, Yao Zhang, Gengyuan Zhang, Jinhe Bi, Philip Torr, Jindong Gu, Denis Krompass, Volker Tresp

♻ ☆ Stateful Large Language Model Serving with Pensieve

Lingfan Yu, Jinkun Lin, Jinyang Li

Large Language Models (LLMs) are wildly popular today and it is important to serve them efficiently. Existing LLM serving systems are stateless across requests. Consequently, when LLMs are used in the common setting of multi-turn conversations, a growing log of the conversation history must be processed alongside any request by the serving system at each turn, resulting in repeated processing. In this paper, we design $Pensieve$, a system optimized for multi-turn conversation LLM serving. $Pensieve$ maintains the conversation state across requests by caching previously processed history to avoid duplicate processing. $Pensieve$'s multi-tier caching strategy can utilize both GPU and CPU memory to efficiently store and retrieve cached data. $Pensieve$ also generalizes the recent PagedAttention kernel to support attention between multiple input tokens with a GPU cache spread over non-contiguous memory. Our evaluation shows that $Pensieve$ can achieve $1.14$-$3.0\times$ the throughput of vLLM and TensorRT-LLM and significantly reduce latency.

♻ ☆ SeBS-Flow: Benchmarking Serverless Cloud Function Workflows

Larissa Schmid, Marcin Copik, Alexandru Calotoiu, Laurin Brandner, Anne Koziolek, Torsten Hoefler

Serverless computing has emerged as a prominent paradigm, with a significant adoption rate among cloud customers. While this model offers advantages such as abstraction from the deployment and resource scheduling, it also poses limitations in handling complex use cases due to the restricted nature of individual functions. Serverless workflows address this limitation by orchestrating multiple functions into a cohesive application. However, existing serverless workflow platforms exhibit significant differences in their programming models and infrastructure, making fair and consistent performance evaluations difficult in practice. To address this gap, we propose the first serverless workflow benchmarking suite SeBS-Flow, providing a platform-agnostic workflow model that enables consistent benchmarking across various platforms. SeBS-Flow includes six real-world application benchmarks and four microbenchmarks representing different computational patterns. We conduct comprehensive evaluations on three major cloud platforms, assessing performance, cost, scalability, and runtime deviations. We make our benchmark suite open-source, enabling rigorous and comparable evaluations of serverless workflows over time.

♻ ☆ LayerKV: Optimizing Large Language Model Serving with Layer-wise KV Cache Management

Yi Xiong, Hao Wu, Changxu Shao, Ziqing Wang, Rui Zhang, Yuhong Guo, Junping Zhao, Ke Zhang, Zhenxuan Pan

The expanding context windows in large language models (LLMs) have greatly enhanced their capabilities in various applications, but they also introduce significant challenges in maintaining low latency, particularly in Time to First Token (TTFT). This paper identifies that the sharp rise in TTFT as context length increases is predominantly driven by queuing delays, which are caused by the growing demands for GPU Key-Value (KV) cache allocation clashing with the limited availability of KV cache blocks. To address this issue, we propose LayerKV, a simple yet effective plug-in method that effectively reduces TTFT without requiring additional hardware or compromising output performance, while seamlessly integrating with existing parallelism strategies and scheduling techniques. Specifically, LayerKV introduces layer-wise KV block allocation, management, and offloading for fine-grained control over system memory, coupled with an SLO-aware scheduler to optimize overall Service Level Objectives (SLOs). Comprehensive evaluations on representative models, ranging from 7B to 70B parameters, across various GPU configurations, demonstrate that LayerKV improves TTFT latency up to 69x and reduces SLO violation rates by 28.7%, significantly enhancing the user experience.

comment: 11 pages, 7 figures, 1 table

♻ ☆ Space-time deterministic graph rewriting

Pablo Arrighi, Marin Costes, Gilles Dowek, Luidnel Maignan

We study non-terminating graph rewriting models, whose local rules are applied non-deterministically -- and yet enjoy a strong form of determinism, namely space-time determinism. Of course in the case of terminating computation it is well-known that the mess introduced by asynchronous rule applications may not matter to the end result, as confluence conspires to produce a unique normal form. In the context of non-terminating computation however, confluence is a very weak property, and (almost) synchronous rule applications is always preferred e.g. when it comes to simulating dynamical systems. Here we provide sufficient conditions so that asynchronous local rule applications conspire to produce well-determined events in the space-time unfolding of the graph, regardless of their application orders. Our first example is an asynchronous simulation of a dynamical system. Our second example features time dilation, in the spirit of general relativity.

♻ ☆ Scalable and Adaptively Secure Any-Trust Distributed Key Generation and All-hands Checkpointing

Hanwen Feng, Tiancheng Mai, Qiang Tang

The classical distributed key generation protocols (DKG) are resurging due to their widespread applications in blockchain. While efforts have been made to improve DKG communication, practical large-scale deployments are still yet to come due to various challenges, including the heavy computation and communication (particularly broadcast) overhead in their adversarial cases. In this paper, we propose a practical DKG for DLog-based cryptosystems, which achieves (quasi-)linear computation and communication per-node cost with the help of a common coin, even in the face of the maximal amount of Byzantine nodes. Moreover, our protocol is secure against adaptive adversaries, which can corrupt less than half of all nodes. The key to our improvements lies in delegating the most costly operations to an Any-Trust group together with a set of techniques for adaptive security. This group is randomly sampled and consists of a small number of individuals. The population only trusts that at least one member in the group is honest, without knowing which one. Moreover, we present a generic transformer that enables us to efficiently deploy a conventional distributed protocol like our DKG, even when the participants have different weights. Additionally, we introduce an extended broadcast channel based on a blockchain and data dispersal network (such as IPFS), enabling reliable broadcasting of arbitrary-size messages at the cost of constant-size blockchain storage.

comment: 22 pages, 5 figures

Machine Learning 151

☆ Data Advisor: Dynamic Data Curation for Safety Alignment of Large Language Models EMNLP 2024

Fei Wang, Ninareh Mehrabi, Palash Goyal, Rahul Gupta, Kai-Wei Chang, Aram Galstyan

comment: Accepted to EMNLP 2024 Main Conference. Project website: https://feiwang96.github.io/DataAdvisor/

☆ PrefixQuant: Static Quantization Beats Dynamic through Prefixed Outliers in LLMs

Mengzhao Chen, Yi Liu, Jiahao Wang, Yi Bin, Wenqi Shao, Ping Luo

comment: A PTQ method to significantly boost the performance of static activation quantization

☆ Regression Conformal Prediction under Bias

Matt Y. Cheung, Tucker J. Netherton, Laurence E. Court, Ashok Veeraraghavan, Guha Balakrishnan

Uncertainty quantification is crucial to account for the imperfect predictions of machine learning algorithms for high-impact applications. Conformal prediction (CP) is a powerful framework for uncertainty quantification that generates calibrated prediction intervals with valid coverage. In this work, we study how CP intervals are affected by bias - the systematic deviation of a prediction from ground truth values - a phenomenon prevalent in many real-world applications. We investigate the influence of bias on interval lengths of two different types of adjustments -- symmetric adjustments, the conventional method where both sides of the interval are adjusted equally, and asymmetric adjustments, a more flexible method where the interval can be adjusted unequally in positive or negative directions. We present theoretical and empirical analyses characterizing how symmetric and asymmetric adjustments impact the "tightness" of CP intervals for regression tasks. Specifically for absolute residual and quantile-based non-conformity scores, we prove: 1) the upper bound of symmetrically adjusted interval lengths increases by $2|b|$ where $b$ is a globally applied scalar value representing bias, 2) asymmetrically adjusted interval lengths are not affected by bias, and 3) conditions when asymmetrically adjusted interval lengths are guaranteed to be smaller than symmetric ones. Our analyses suggest that even if predictions exhibit significant drift from ground truth values, asymmetrically adjusted intervals are still able to maintain the same tightness and validity of intervals as if the drift had never happened, while symmetric ones significantly inflate the lengths. We demonstrate our theoretical results with two real-world prediction tasks: sparse-view computed tomography (CT) reconstruction and time-series weather forecasting. Our work paves the way for more bias-robust machine learning systems.

comment: 17 pages, 6 figures, code available at: https://github.com/matthewyccheung/conformal-metric

☆ Differential Transformer

Tianzhu Ye, Li Dong, Yuqing Xia, Yutao Sun, Yi Zhu, Gao Huang, Furu Wei

☆ SePPO: Semi-Policy Preference Optimization for Diffusion Alignment

Daoan Zhang, Guangchen Lan, Dong-Jun Han, Wenlin Yao, Xiaoman Pan, Hongming Zhang, Mingxiao Li, Pengcheng Chen, Yu Dong, Christopher Brinton, Jiebo Luo

☆ GLEE: A Unified Framework and Benchmark for Language-based Economic Environments

Eilam Shapira, Omer Madmon, Itamar Reinman, Samuel Joseph Amouyal, Roi Reichart, Moshe Tennenholtz

☆ Causal Micro-Narratives EMNLP 2024

Mourad Heddaya, Qingcheng Zeng, Chenhao Tan, Rob Voigt, Alexander Zentefis

comment: Accepted to EMNLP 2024 Workshop on Narrative Understanding

☆ SFTMix: Elevating Language Model Instruction Tuning with Mixup Recipe

Yuxin Xiao, Shujian Zhang, Wenxuan Zhou, Marzyeh Ghassemi, Sanqiang Zhao

☆ SimO Loss: Anchor-Free Contrastive Loss for Fine-Grained Supervised Contrastive Learning

Taha Bouhsine, Imad El Aaroussi, Atik Faysal, Wang Huaxia

☆ SymmetryLens: A new candidate paradigm for unsupervised symmetry learning via locality and equivariance

Onur Efe, Arkadas Ozakin

We develop a new, unsupervised symmetry learning method that starts with raw data, and gives the minimal (discrete) generator of an underlying Lie group of symmetries, together with a symmetry equivariant representation of the data. The method is able to learn the pixel translation operator from a dataset with only an approximate translation symmetry, and can learn quite different types of symmetries which are not apparent to the naked eye, equally well. The method is based on the formulation of an information-theoretic loss function that measures both the degree to which the dataset is symmetric under a given candidate symmetry, and also, the degree of locality of the samples in the dataset with respect to this symmetry. We demonstrate that this coupling between symmetry and locality, together with a special optimization technique developed for entropy estimation, results in a highly stable system that gives reproducible results. The symmetry actions we consider are group representations, however, we believe the approach has the potential to be generalized to more general, nonlinear actions of non-commutative Lie groups.

comment: 27 pages

☆ GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models

Iman Mirzadeh, Keivan Alizadeh, Hooman Shahrokhi, Oncel Tuzel, Samy Bengio, Mehrdad Farajtabar

Recent advancements in Large Language Models (LLMs) have sparked interest in their formal reasoning capabilities, particularly in mathematics. The GSM8K benchmark is widely used to assess the mathematical reasoning of models on grade-school-level questions. While the performance of LLMs on GSM8K has significantly improved in recent years, it remains unclear whether their mathematical reasoning capabilities have genuinely advanced, raising questions about the reliability of the reported metrics. To address these concerns, we conduct a large-scale study on several SOTA open and closed models. To overcome the limitations of existing evaluations, we introduce GSM-Symbolic, an improved benchmark created from symbolic templates that allow for the generation of a diverse set of questions. GSM-Symbolic enables more controllable evaluations, providing key insights and more reliable metrics for measuring the reasoning capabilities of models.Our findings reveal that LLMs exhibit noticeable variance when responding to different instantiations of the same question. Specifically, the performance of all models declines when only the numerical values in the question are altered in the GSM-Symbolic benchmark. Furthermore, we investigate the fragility of mathematical reasoning in these models and show that their performance significantly deteriorates as the number of clauses in a question increases. We hypothesize that this decline is because current LLMs cannot perform genuine logical reasoning; they replicate reasoning steps from their training data. Adding a single clause that seems relevant to the question causes significant performance drops (up to 65%) across all state-of-the-art models, even though the clause doesn't contribute to the reasoning chain needed for the final answer. Overall, our work offers a more nuanced understanding of LLMs' capabilities and limitations in mathematical reasoning.

comment: preprint

☆ ETGL-DDPG: A Deep Deterministic Policy Gradient Algorithm for Sparse Reward Continuous Control

Ehsan Futuhi, Shayan Karimi, Chao Gao, Martin Müller

We consider deep deterministic policy gradient (DDPG) in the context of reinforcement learning with sparse rewards. To enhance exploration, we introduce a search procedure, \emph{${\epsilon}{t}$-greedy}, which generates exploratory options for exploring less-visited states. We prove that search using $\epsilon t$-greedy has polynomial sample complexity under mild MDP assumptions. To more efficiently use the information provided by rewarded transitions, we develop a new dual experience replay buffer framework, \emph{GDRB}, and implement \emph{longest n-step returns}. The resulting algorithm, \emph{ETGL-DDPG}, integrates all three techniques: \bm{$\epsilon t$}-greedy, \textbf{G}DRB, and \textbf{L}ongest $n$-step, into DDPG. We evaluate ETGL-DDPG on standard benchmarks and demonstrate that it outperforms DDPG, as well as other state-of-the-art methods, across all tested sparse-reward continuous environments. Ablation studies further highlight how each strategy individually enhances the performance of DDPG in this setting.

☆ Cookbook: A framework for improving LLM generative abilities via programmatic data generating templates

Avanika Narayan, Mayee F. Chen, Kush Bhatia, Christopher Ré

comment: COLM 2024

☆ Precise Model Benchmarking with Only a Few Observations EMNLP 2024

Riccardo Fogliato, Pratik Patil, Nil-Jana Akpinar, Mathew Monfort

comment: To appear at EMNLP 2024

☆ Density estimation with LLMs: a geometric investigation of in-context learning trajectories ICLR 2025

Toni J. B. Liu, Nicolas Boullé, Raphaël Sarfati, Christopher J. Earls

comment: Under review as a conference paper at ICLR 2025

☆ Beyond FVD: Enhanced Evaluation Metrics for Video Generation Quality

Ge Ya, Luo, Gian Favero, Zhi Hao Luo, Alexia Jolicoeur-Martineau, Christopher Pal

☆ Understanding Warmup-Stable-Decay Learning Rates: A River Valley Loss Landscape Perspective

Kaiyue Wen, Zhiyuan Li, Jason Wang, David Hall, Percy Liang, Tengyu Ma

comment: 45 pages,13 figures

☆ Matrix-weighted networks for modeling multidimensional dynamics

Yu Tian, Sadamori Kojaku, Hiroki Sayama, Renaud Lambiotte

Networks are powerful tools for modeling interactions in complex systems. While traditional networks use scalar edge weights, many real-world systems involve multidimensional interactions. For example, in social networks, individuals often have multiple interconnected opinions that can affect different opinions of other individuals, which can be better characterized by matrices. We propose a novel, general framework for modeling such multidimensional interacting dynamics: matrix-weighted networks (MWNs). We present the mathematical foundations of MWNs and examine consensus dynamics and random walks within this context. Our results reveal that the coherence of MWNs gives rise to non-trivial steady states that generalize the notions of communities and structural balance in traditional networks.

comment: 14 pages, 8 figures

☆ MARs: Multi-view Attention Regularizations for Patch-based Feature Recognition of Space Terrain ECCV 2024

Timothy Chase Jr, Karthik Dantu

comment: ECCV 2024. Project page available at https://droneslab.github.io/mars/

☆ Are causal effect estimations enough for optimal recommendations under multitreatment scenarios?

Sherly Alfonso-Sánchez, Kristina P. Sendova, Cristián Bravo

When making treatment selection decisions, it is essential to include a causal effect estimation analysis to compare potential outcomes under different treatments or controls, assisting in optimal selection. However, merely estimating individual treatment effects may not suffice for truly optimal decisions. Our study addressed this issue by incorporating additional criteria, such as the estimations' uncertainty, measured by the conditional value-at-risk, commonly used in portfolio and insurance management. For continuous outcomes observable before and after treatment, we incorporated a specific prediction condition. We prioritized treatments that could yield optimal treatment effect results and lead to post-treatment outcomes more desirable than pretreatment levels, with the latter condition being called the prediction criterion. With these considerations, we propose a comprehensive methodology for multitreatment selection. Our approach ensures satisfaction of the overlap assumption, crucial for comparing outcomes for treated and control groups, by training propensity score models as a preliminary step before employing traditional causal models. To illustrate a practical application of our methodology, we applied it to the credit card limit adjustment problem. Analyzing a fintech company's historical data, we found that relying solely on counterfactual predictions was inadequate for appropriate credit line modifications. Incorporating our proposed additional criteria significantly enhanced policy performance.

comment: 34 pages, 4 figures

☆ Presto! Distilling Steps and Layers for Accelerating Music Generation

Zachary Novack, Ge Zhu, Jonah Casebeer, Julian McAuley, Taylor Berg-Kirkpatrick, Nicholas J. Bryan

Despite advances in diffusion-based text-to-music (TTM) methods, efficient, high-quality generation remains a challenge. We introduce Presto!, an approach to inference acceleration for score-based diffusion transformers via reducing both sampling steps and cost per step. To reduce steps, we develop a new score-based distribution matching distillation (DMD) method for the EDM-family of diffusion models, the first GAN-based distillation method for TTM. To reduce the cost per step, we develop a simple, but powerful improvement to a recent layer distillation method that improves learning via better preserving hidden state variance. Finally, we combine our step and layer distillation methods together for a dual-faceted approach. We evaluate our step and layer distillation methods independently and show each yield best-in-class performance. Our combined distillation method can generate high-quality outputs with improved diversity, accelerating our base model by 10-18x (230/435ms latency for 32 second mono/stereo 44.1kHz, 15x faster than comparable SOTA) -- the fastest high-quality TTM to our knowledge. Sound examples can be found at https://presto-music.github.io/web/.

☆ A Simulation-Free Deep Learning Approach to Stochastic Optimal Control

Mengjian Hua, Matthieu Laurière, Eric Vanden-Eijnden

☆ PAMLR: A Passive-Active Multi-Armed Bandit-Based Solution for LoRa Channel Allocation

Jihoon Yun, Chengzhang Li, Anish Arora

comment: 10 pages

☆ Tuning-Free Bilevel Optimization: New Algorithms and Convergence Analysis

Yifan Yang, Hao Ban, Minhui Huang, Shiqian Ma, Kaiyi Ji

Bilevel optimization has recently attracted considerable attention due to its abundant applications in machine learning problems. However, existing methods rely on prior knowledge of problem parameters to determine stepsizes, resulting in significant effort in tuning stepsizes when these parameters are unknown. In this paper, we propose two novel tuning-free algorithms, D-TFBO and S-TFBO. D-TFBO employs a double-loop structure with stepsizes adaptively adjusted by the "inverse of cumulative gradient norms" strategy. S-TFBO features a simpler fully single-loop structure that updates three variables simultaneously with a theory-motivated joint design of adaptive stepsizes for all variables. We provide a comprehensive convergence analysis for both algorithms and show that D-TFBO and S-TFBO respectively require $O(\frac{1}{\epsilon})$ and $O(\frac{1}{\epsilon}\log^4(\frac{1}{\epsilon}))$ iterations to find an $\epsilon$-accurate stationary point, (nearly) matching their well-tuned counterparts using the information of problem parameters. Experiments on various problems show that our methods achieve performance comparable to existing well-tuned approaches, while being more robust to the selection of initial stepsizes. To the best of our knowledge, our methods are the first to completely eliminate the need for stepsize tuning, while achieving theoretical guarantees.

☆ LOTOS: Layer-wise Orthogonalization for Training Robust Ensembles

Ali Ebrahimpour-Boroojeny, Hari Sundaram, Varun Chandrasekaran

Transferability of adversarial examples is a well-known property that endangers all classification models, even those that are only accessible through black-box queries. Prior work has shown that an ensemble of models is more resilient to transferability: the probability that an adversarial example is effective against most models of the ensemble is low. Thus, most ongoing research focuses on improving ensemble diversity. Another line of prior work has shown that Lipschitz continuity of the models can make models more robust since it limits how a model's output changes with small input perturbations. In this paper, we study the effect of Lipschitz continuity on transferability rates. We show that although a lower Lipschitz constant increases the robustness of a single model, it is not as beneficial in training robust ensembles as it increases the transferability rate of adversarial examples across models in the ensemble. Therefore, we introduce LOTOS, a new training paradigm for ensembles, which counteracts this adverse effect. It does so by promoting orthogonality among the top-$k$ sub-spaces of the transformations of the corresponding affine layers of any pair of models in the ensemble. We theoretically show that $k$ does not need to be large for convolutional layers, which makes the computational overhead negligible. Through various experiments, we show LOTOS increases the robust accuracy of ensembles of ResNet-18 models by $6$ percentage points (p.p) against black-box attacks on CIFAR-10. It is also capable of combining with the robustness of prior state-of-the-art methods for training robust ensembles to enhance their robust accuracy by $10.7$ p.p.

☆ A Digital Twin Framework for Liquid-cooled Supercomputers as Demonstrated at Exascale

Wesley Brewer, Matthias Maiterth, Vineet Kumar, Rafal Wojda, Sedrick Bouknight, Jesse Hines, Woong Shin, Scott Greenwood, David Grant, Wesley Williams, Feiyi Wang

comment: 14 pages, 9 figures, To be published in the Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis. 2024

☆ Agnostic Smoothed Online Learning

Moïse Blanchard

Classical results in statistical learning typically consider two extreme data-generating models: i.i.d. instances from an unknown distribution, or fully adversarial instances, often much more challenging statistically. To bridge the gap between these models, recent work introduced the smoothed framework, in which at each iteration an adversary generates instances from a distribution constrained to have density bounded by $\sigma^{-1}$ compared to some fixed base measure $\mu$. This framework interpolates between the i.i.d. and adversarial cases, depending on the value of $\sigma$. For the classical online prediction problem, most prior results in smoothed online learning rely on the arguably strong assumption that the base measure $\mu$ is known to the learner, contrasting with standard settings in the PAC learning or consistency literature. We consider the general agnostic problem in which the base measure is unknown and values are arbitrary. Along this direction, Block et al. showed that empirical risk minimization has sublinear regret under the well-specified assumption. We propose an algorithm R-Cover based on recursive coverings which is the first to guarantee sublinear regret for agnostic smoothed online learning without prior knowledge of $\mu$. For classification, we prove that R-Cover has adaptive regret $\tilde O(\sqrt{dT/\sigma})$ for function classes with VC dimension $d$, which is optimal up to logarithmic factors. For regression, we establish that R-Cover has sublinear oblivious regret for function classes with polynomial fat-shattering dimension growth.

☆ Assouad, Fano, and Le Cam with Interaction: A Unifying Lower Bound Framework and Characterization for Bandit Learnability

Fan Chen, Dylan J. Foster, Yanjun Han, Jian Qian, Alexander Rakhlin, Yunbei Xu

In this paper, we develop a unified framework for lower bound methods in statistical estimation and interactive decision making. Classical lower bound techniques -- such as Fano's inequality, Le Cam's method, and Assouad's lemma -- have been central to the study of minimax risk in statistical estimation, yet they are insufficient for the analysis of methods that collect data in an interactive manner. The recent minimax lower bounds for interactive decision making via the Decision-Estimation Coefficient (DEC) appear to be genuinely different from the classical methods. We propose a unified view of these distinct methodologies through a general algorithmic lower bound method. We further introduce a novel complexity measure, decision dimension, which facilitates the derivation of new lower bounds for interactive decision making. In particular, decision dimension provides a characterization of bandit learnability for any structured bandit model class. Further, we characterize the sample complexity of learning convex model class up to a polynomial gap with the decision dimension, addressing the remaining gap between upper and lower bounds in Foster et al. (2021, 2023).

☆ Human-Feedback Efficient Reinforcement Learning for Online Diffusion Model Finetuning

Ayano Hiranaka, Shang-Fu Chen, Chieh-Hsin Lai, Dongjun Kim, Naoki Murata, Takashi Shibuya, Wei-Hsiang Liao, Shao-Hua Sun, Yuki Mitsufuji

☆ Hyper-Representations: Learning from Populations of Neural Networks

Konstantin Schürholt

This thesis addresses the challenge of understanding Neural Networks through the lens of their most fundamental component: the weights, which encapsulate the learned information and determine the model behavior. At the core of this thesis is a fundamental question: Can we learn general, task-agnostic representations from populations of Neural Network models? The key contribution of this thesis to answer that question are hyper-representations, a self-supervised method to learn representations of NN weights. Work in this thesis finds that trained NN models indeed occupy meaningful structures in the weight space, that can be learned and used. Through extensive experiments, this thesis demonstrates that hyper-representations uncover model properties, such as their performance, state of training, or hyperparameters. Moreover, the identification of regions with specific properties in hyper-representation space allows to sample and generate model weights with targeted properties. This thesis demonstrates applications for fine-tuning, and transfer learning to great success. Lastly, it presents methods that allow hyper-representations to generalize beyond model sizes, architectures, and tasks. The practical implications of that are profound, as it opens the door to foundation models of Neural Networks, which aggregate and instantiate their knowledge across models and architectures. Ultimately, this thesis contributes to the deeper understanding of Neural Networks by investigating structures in their weights which leads to more interpretable, efficient, and adaptable models. By laying the groundwork for representation learning of NN weights, this research demonstrates the potential to change the way Neural Networks are developed, analyzed, and used.

comment: PhD Dissertation accepted at University of St. Gallen

☆ Nonasymptotic Analysis of Stochastic Gradient Descent with the Richardson-Romberg Extrapolation

Marina Sheshukova, Denis Belomestny, Alain Durmus, Eric Moulines, Alexey Naumov, Sergey Samsonov

We address the problem of solving strongly convex and smooth minimization problems using stochastic gradient descent (SGD) algorithm with a constant step size. Previous works suggested to combine the Polyak-Ruppert averaging procedure with the Richardson-Romberg extrapolation technique to reduce the asymptotic bias of SGD at the expense of a mild increase of the variance. We significantly extend previous results by providing an expansion of the mean-squared error of the resulting estimator with respect to the number of iterations $n$. More precisely, we show that the mean-squared error can be decomposed into the sum of two terms: a leading one of order $\mathcal{O}(n^{-1/2})$ with explicit dependence on a minimax-optimal asymptotic covariance matrix, and a second-order term of order $\mathcal{O}(n^{-3/4})$ where the power $3/4$ can not be improved in general. We also extend this result to the $p$-th moment bound keeping optimal scaling of the remainders with respect to $n$. Our analysis relies on the properties of the SGD iterates viewed as a time-homogeneous Markov chain. In particular, we establish that this chain is geometrically ergodic with respect to a suitably defined weighted Wasserstein semimetric.

☆ SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Masks

Fenia Christopoulou, Ronald Cardenas, Gerasimos Lampouras, Haitham Bou-Ammar, Jun Wang

comment: 20 papges, 9 figures, 5 tables. Under Review

☆ CR-CTC: Consistency regularization on CTC for improved speech recognition

Zengwei Yao, Wei Kang, Xiaoyu Yang, Fangjun Kuang, Liyong Guo, Han Zhu, Zengrui Jin, Zhaoqing Li, Long Lin, Daniel Povey

Connectionist Temporal Classification (CTC) is a widely used method for automatic speech recognition (ASR), renowned for its simplicity and computational efficiency. However, it often falls short in recognition performance compared to transducer or systems combining CTC and attention-based encoder-decoder (CTC/AED). In this work, we propose the Consistency-Regularized CTC (CR-CTC), which enforces consistency between two CTC distributions obtained from different augmented views of the input speech mel-spectrogram. We provide in-depth insights into its essential behaviors from three perspectives: 1) it conducts self-distillation between random pairs of sub-models that process different augmented views; 2) it learns contextual representation through masked prediction for positions within time-masked regions, especially when we increase the amount of time masking; 3) it suppresses the extremely peaky CTC distributions, thereby reducing overfitting and improving the generalization ability. Extensive experiments on LibriSpeech, Aishell-1, and GigaSpeech datasets demonstrate the effectiveness of our CR-CTC, which achieves performance comparable to, or even slightly better than, that of transducer and CTC/AED.

☆ DreamSat: Towards a General 3D Model for Novel View Synthesis of Space Objects

Nidhi Mathihalli, Audrey Wei, Giovanni Lavezzi, Peng Mun Siew, Victor Rodriguez-Fernandez, Hodei Urrutxua, Richard Linares

comment: Presented at the 75th International Astronautical Congress, October 2024, Milan, Italy

☆ HyperINF: Unleashing the HyperPower of the Schulz's Method for Data Influence Estimation

Xinyu Zhou, Simin Fan, Martin Jaggi

Influence functions provide a principled method to assess the contribution of individual training samples to a specific target. Yet, their high computational costs limit their applications on large-scale models and datasets. Existing methods proposed for influence function approximation have significantly reduced the computational overheads. However, they mostly suffer from inaccurate estimation due to the lack of strong convergence guarantees from the algorithm. The family of hyperpower methods are well-known for their rigorous convergence guarantees on matrix inverse approximation, while the matrix multiplication operation can involve intractable memory and computation costs on large-scale models. We propose HyperINF, an efficient and accurate influence function approximation method which leverages the hyperpower method, specifically Schulz's iterative algorithm. To deal with the computation-intensive matrix multiplication, we incorporate the generalized fisher information (GFIM) as a low-rank approximation of the Hessian matrix, which reduces the memory and computation overheads to constant costs independent of ranks on LoRA-tuned models. We first demonstrate the superior accuracy and stability of \method compared to other baselines through a synthetic convergence simulation for matrix inversion. We further validate the efficacy of \method through extensive real-world data attribution tasks, including mislabeled data detection and data selection for LLM and VLM fine-tuning. On LoRA-tuned models, HyperINF achieves superior downstream performance with minimal memory and computational overhead, while other baselines suffer from significant degradation. Our codebase is available at https://github.com/Blackzxy/HyperINF.

☆ ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery

comment: 55 pages

☆ Compression via Pre-trained Transformers: A Study on Byte-Level Multimodal Data

David Heurtel-Depeiges, Anian Ruoss, Joel Veness, Tim Genewein

Foundation models have recently been shown to be strong data compressors. However, when accounting for their excessive parameter count, their compression ratios are actually inferior to standard compression algorithms. Moreover, naively reducing the number of parameters may not necessarily help as it leads to worse predictions and thus weaker compression. In this paper, we conduct a large-scale empirical study to investigate whether there is a sweet spot where competitive compression ratios with pre-trained vanilla transformers are possible. To this end, we train families of models on 165GB of raw byte sequences of either text, image, or audio data (and all possible combinations of the three) and then compress 1GB of out-of-distribution (OOD) data from each modality. We find that relatively small models (i.e., millions of parameters) can outperform standard general-purpose compression algorithms (gzip, LZMA2) and even domain-specific compressors (PNG, JPEG 2000, FLAC) - even when factoring in parameter count. We achieve, e.g., the lowest compression ratio of 0.49 on OOD audio data (vs. 0.54 for FLAC). To study the impact of model- and dataset scale, we conduct extensive ablations and hyperparameter sweeps, and we investigate the effect of unimodal versus multimodal training. We find that even small models can be trained to perform well on multiple modalities, but, in contrast to previously reported results with large-scale foundation models, transfer to unseen modalities is generally weak.

☆ TidalDecode: Fast and Accurate LLM Decoding with Position Persistent Sparse Attention

Lijie Yang, Zhihao Zhang, Zhuofu Chen, Zikun Li, Zhihao Jia

☆ Function Gradient Approximation with Random Shallow ReLU Networks with Control Applications

Andrew Lamperski, Siddharth Salapaka

Neural networks are widely used to approximate unknown functions in control. A common neural network architecture uses a single hidden layer (i.e. a shallow network), in which the input parameters are fixed in advance and only the output parameters are trained. The typical formal analysis asserts that if output parameters exist to approximate the unknown function with sufficient accuracy, then desired control performance can be achieved. A long-standing theoretical gap was that no conditions existed to guarantee that, for the fixed input parameters, required accuracy could be obtained by training the output parameters. Our recent work has partially closed this gap by demonstrating that if input parameters are chosen randomly, then for any sufficiently smooth function, with high-probability there are output parameters resulting in $O((1/m)^{1/2})$ approximation errors, where $m$ is the number of neurons. However, some applications, notably continuous-time value function approximation, require that the network approximates the both the unknown function and its gradient with sufficient accuracy. In this paper, we show that randomly generated input parameters and trained output parameters result in gradient errors of $O((\log(m)/m)^{1/2})$, and additionally, improve the constants from our prior work. We show how to apply the result to policy evaluation problems.

comment: Under Review for American Control Conference, 2025

☆ Control-oriented Clustering of Visual Latent Representation

Han Qi, Haocheng Yin, Heng Yang

☆ SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification NeurIPS 2024

Benjamin Feuer, Jiawei Xu, Niv Cohen, Patrick Yubeaton, Govind Mittal, Chinmay Hegde

comment: NeurIPS 2024, Datasets and Benchmarks Track

☆ FreSh: Frequency Shifting for Accelerated Neural Representation Learning

Adam Kania, Marko Mihajlovic, Sergey Prokudin, Jacek Tabor, Przemysław Spurek

☆ PhotoReg: Photometrically Registering 3D Gaussian Splatting Models

Ziwen Yuan, Tianyi Zhang, Matthew Johnson-Roberson, Weiming Zhi

☆ Systematic Literature Review of Vision-Based Approaches to Outdoor Livestock Monitoring with Lessons from Wildlife Studies

Stacey D. Scott, Zayn J. Abbas, Feerass Ellid, Eli-Henry Dykhne, Muhammad Muhaiminul Islam, Weam Ayad, Kristina Kacmorova, Dan Tulpan, Minglun Gong

comment: 28 pages, 5 figures, 2 tables

☆ Active Fine-Tuning of Generalist Policies

Marco Bagatella, Jonas Hübotter, Georg Martius, Andreas Krause

Pre-trained generalist policies are rapidly gaining relevance in robot learning due to their promise of fast adaptation to novel, in-domain tasks. This adaptation often relies on collecting new demonstrations for a specific task of interest and applying imitation learning algorithms, such as behavioral cloning. However, as soon as several tasks need to be learned, we must decide which tasks should be demonstrated and how often? We study this multi-task problem and explore an interactive framework in which the agent adaptively selects the tasks to be demonstrated. We propose AMF (Active Multi-task Fine-tuning), an algorithm to maximize multi-task policy performance under a limited demonstration budget by collecting demonstrations yielding the largest information gain on the expert policy. We derive performance guarantees for AMF under regularity assumptions and demonstrate its empirical effectiveness to efficiently fine-tune neural policies in complex and high-dimensional environments.

☆ DEPT: Decoupled Embeddings for Pre-training Language Models

Alex Iacob, Lorenzo Sani, Meghdad Kurmanji, William F. Shen, Xinchi Qiu, Dongqi Cai, Yan Gao, Nicholas D. Lane

☆ FRIDA: Free-Rider Detection using Privacy Attacks

Pol G. Recasens, Ádám Horváth, Alberto Gutierrez-Torre, Jordi Torres, Josep Ll. Berral, Balázs Pejó

Federated learning is increasingly popular as it enables multiple parties with limited datasets and resources to train a high-performing machine learning model collaboratively. However, similarly to other collaborative systems, federated learning is vulnerable to free-riders -- participants who do not contribute to the training but still benefit from the shared model. Free-riders not only compromise the integrity of the learning process but also slow down the convergence of the global model, resulting in increased costs for the honest participants. To address this challenge, we propose FRIDA: free-rider detection using privacy attacks, a framework that leverages inference attacks to detect free-riders. Unlike traditional methods that only capture the implicit effects of free-riding, FRIDA directly infers details of the underlying training datasets, revealing characteristics that indicate free-rider behaviour. Through extensive experiments, we demonstrate that membership and property inference attacks are effective for this purpose. Our evaluation shows that FRIDA outperforms state-of-the-art methods, especially in non-IID settings.

☆ RelUNet: Relative Channel Fusion U-Net for Multichannel Speech Enhancement

Ibrahim Aldarmaki, Thamar Solorio, Bhiksha Raj, Hanan Aldarmaki

Neural multi-channel speech enhancement models, in particular those based on the U-Net architecture, demonstrate promising performance and generalization potential. These models typically encode input channels independently, and integrate the channels during later stages of the network. In this paper, we propose a novel modification of these models by incorporating relative information from the outset, where each channel is processed in conjunction with a reference channel through stacking. This input strategy exploits comparative differences to adaptively fuse information between channels, thereby capturing crucial spatial information and enhancing the overall performance. The experiments conducted on the CHiME-3 dataset demonstrate improvements in speech enhancement metrics across various architectures.

☆ T-JEPA: Augmentation-Free Self-Supervised Learning for Tabular Data

Hugo Thimonier, José Lucas De Melo Costa, Fabrice Popineau, Arpad Rimmel, Bich-Liên Doan

Self-supervision is often used for pre-training to foster performance on a downstream task by constructing meaningful representations of samples. Self-supervised learning (SSL) generally involves generating different views of the same sample and thus requires data augmentations that are challenging to construct for tabular data. This constitutes one of the main challenges of self-supervision for structured data. In the present work, we propose a novel augmentation-free SSL method for tabular data. Our approach, T-JEPA, relies on a Joint Embedding Predictive Architecture (JEPA) and is akin to mask reconstruction in the latent space. It involves predicting the latent representation of one subset of features from the latent representation of a different subset within the same sample, thereby learning rich representations without augmentations. We use our method as a pre-training technique and train several deep classifiers on the obtained representation. Our experimental results demonstrate a substantial improvement in both classification and regression tasks, outperforming models trained directly on samples in their original data space. Moreover, T-JEPA enables some methods to consistently outperform or match the performance of traditional methods likes Gradient Boosted Decision Trees. To understand why, we extensively characterize the obtained representations and show that T-JEPA effectively identifies relevant features for downstream tasks without access to the labels. Additionally, we introduce regularization tokens, a novel regularization method critical for training of JEPA-based models on structured data.

☆ Assumption-Lean Post-Integrated Inference with Negative Control Outcomes

Jin-Hong Du, Kathryn Roeder, Larry Wasserman

Data integration has become increasingly common in aligning multiple heterogeneous datasets. With high-dimensional outcomes, data integration methods aim to extract low-dimensional embeddings of observations to remove unwanted variations, such as batch effects and unmeasured covariates, inherent in data collected from different sources. However, multiple hypothesis testing after data integration can be substantially biased due to the data-dependent integration processes. To address this challenge, we introduce a robust post-integrated inference (PII) method that adjusts for latent heterogeneity using negative control outcomes. By leveraging causal interpretations, we derive nonparametric identification conditions that form the basis of our PII approach. Our assumption-lean semiparametric inference method extends robustness and generality to projected direct effect estimands that account for mediators, confounders, and moderators. These estimands remain statistically meaningful under model misspecifications and with error-prone embeddings. We provide deterministic quantifications of the bias of target estimands induced by estimated embeddings and finite-sample linear expansions of the estimators with uniform concentration bounds on the residuals for all outcomes. The proposed doubly robust estimators are consistent and efficient under minimal assumptions, facilitating data-adaptive estimation with machine learning algorithms. Using random forests, we evaluate empirical statistical errors in simulations and analyze single-cell CRISPR perturbed datasets with potential unmeasured confounders.

comment: 29 pages for main text, and 18 pages for appendix, 9 figures for main text, 4 figures for appendix

☆ MC-QDSNN: Quantized Deep evolutionary SNN with Multi-Dendritic Compartment Neurons for Stress Detection using Physiological Signals

Ajay B. S., Phani Pavan K, Madhav Rao

comment: 13 pages, 15 figures. Applied to IEEE Transactions on Computer Aided Design Journal. Awaiting a verdict

☆ Efficient Model-Based Reinforcement Learning Through Optimistic Thompson Sampling

Jasmine Bayrooti, Carl Henrik Ek, Amanda Prorok

Learning complex robot behavior through interactions with the environment necessitates principled exploration. Effective strategies should prioritize exploring regions of the state-action space that maximize rewards, with optimistic exploration emerging as a promising direction aligned with this idea and enabling sample-efficient reinforcement learning. However, existing methods overlook a crucial aspect: the need for optimism to be informed by a belief connecting the reward and state. To address this, we propose a practical, theoretically grounded approach to optimistic exploration based on Thompson sampling. Our model structure is the first that allows for reasoning about joint uncertainty over transitions and rewards. We apply our method on a set of MuJoCo and VMAS continuous control tasks. Our experiments demonstrate that optimistic exploration significantly accelerates learning in environments with sparse rewards, action penalties, and difficult-to-explore regions. Furthermore, we provide insights into when optimism is beneficial and emphasize the critical role of model uncertainty in guiding exploration.

☆ Safe Learning-Based Optimization of Model Predictive Control: Application to Battery Fast-Charging

Sebastian Hirt, Andreas Höhl, Johannes Pohlodek, Joachim Schaeffer, Maik Pfefferkorn, Richard D. Braatz, Rolf Findeisen

Model predictive control (MPC) is a powerful tool for controlling complex nonlinear systems under constraints, but often struggles with model uncertainties and the design of suitable cost functions. To address these challenges, we discuss an approach that integrates MPC with safe Bayesian optimization to optimize long-term closed-loop performance despite significant model-plant mismatches. By parameterizing the MPC stage cost function using a radial basis function network, we employ Bayesian optimization as a multi-episode learning strategy to tune the controller without relying on precise system models. This method mitigates conservativeness introduced by overly cautious soft constraints in the MPC cost function and provides probabilistic safety guarantees during learning, ensuring that safety-critical constraints are met with high probability. As a practical application, we apply our approach to fast charging of lithium-ion batteries, a challenging task due to the complicated battery dynamics and strict safety requirements, subject to the requirement to be implementable in real time. Simulation results demonstrate that, in the context of model-plant mismatch, our method reduces charging times compared to traditional MPC methods while maintaining safety. This work extends previous research by emphasizing closed-loop constraint satisfaction and offers a promising solution for enhancing performance in systems where model uncertainties and safety are critical concerns.

comment: 7 pages, 4 figures, submitted to ACC 2025

☆ Collaboration! Towards Robust Neural Methods for Routing Problems NeurIPS 2024

Jianan Zhou, Yaoxin Wu, Zhiguang Cao, Wen Song, Jie Zhang, Zhiqi Shen

Despite enjoying desirable efficiency and reduced reliance on domain expertise, existing neural methods for vehicle routing problems (VRPs) suffer from severe robustness issues -- their performance significantly deteriorates on clean instances with crafted perturbations. To enhance robustness, we propose an ensemble-based Collaborative Neural Framework (CNF) w.r.t. the defense of neural VRP methods, which is crucial yet underexplored in the literature. Given a neural VRP method, we adversarially train multiple models in a collaborative manner to synergistically promote robustness against attacks, while boosting standard generalization on clean instances. A neural router is designed to adeptly distribute training instances among models, enhancing overall load balancing and collaborative efficacy. Extensive experiments verify the effectiveness and versatility of CNF in defending against various attacks across different neural VRP methods. Notably, our approach also achieves impressive out-of-distribution generalization on benchmark instances.

comment: Accepted at NeurIPS 2024

☆ Failure-Proof Non-Contrastive Self-Supervised Learning

Emanuele Sansone, Tim Lebailly, Tinne Tuytelaars

☆ Detecting and Approximating Redundant Computational Blocks in Neural Networks

Irene Cannistraci, Emanuele Rodolà, Bastian Rieck

comment: 9 pages, 10 figures, 7 tables

☆ Next state prediction gives rise to entangled, yet compositional representations of objects

Tankred Saanum, Luca M. Schulze Buschoff, Peter Dayan, Eric Schulz

☆ Goal-Conditioned Terminal Value Estimation for Real-time and Multi-task Model Predictive Control

Mitsuki Morita, Satoshi Yamamori, Satoshi Yagi, Norikazu Sugimoto, Jun Morimoto

comment: 16 pages, 9 figures

☆ Defense-as-a-Service: Black-box Shielding against Backdoored Graph Models

Xiao Yang, Kai Zhou, Yuni Lai, Gaolei Li

With the trend of large graph learning models, business owners tend to employ a model provided by a third party to deliver business services to users. However, these models might be backdoored, and malicious users can submit trigger-embedded inputs to manipulate the model predictions. Current graph backdoor defenses have several limitations: 1) depending on model-related details, 2) requiring additional model fine-tuning, and 3) relying upon extra explainability tools, all of which are infeasible under stringent privacy policies. To address those limitations, we propose GraphProt, which allows resource-constrained business owners to rely on third parties to avoid backdoor attacks on GNN-based graph classifiers. Our GraphProt is model-agnostic and only relies on the input graph. The key insight is to leverage subgraph information for prediction, thereby mitigating backdoor effects induced by triggers. GraphProt comprises two components: clustering-based trigger elimination and robust subgraph ensemble. Specifically, we first propose feature-topology clustering that aims to remove most of the anomalous subgraphs (triggers). Moreover, we design subgraph sampling strategies based on feature-topology clustering to build a robust classifier via majority vote. Experimental results across three backdoor attacks and six benchmark datasets demonstrate that GraphProt significantly reduces the backdoor attack success rate while preserving the model accuracy on regular graph classification tasks.

☆ Decomposition Polyhedra of Piecewise Linear Functions

Marie-Charlotte Brandenburg, Moritz Grillo, Christoph Hertrich

In this paper we contribute to the frequently studied question of how to decompose a continuous piecewise linear (CPWL) function into a difference of two convex CPWL functions. Every CPWL function has infinitely many such decompositions, but for applications in optimization and neural network theory, it is crucial to find decompositions with as few linear pieces as possible. This is a highly challenging problem, as we further demonstrate by disproving a recently proposed approach by Tran and Wang [Minimal representations of tropical rational functions. Algebraic Statistics, 15(1):27-59, 2024]. To make the problem more tractable, we propose to fix an underlying polyhedral complex determining the possible locus of nonlinearity. Under this assumption, we prove that the set of decompositions forms a polyhedron that arises as intersection of two translated cones. We prove that irreducible decompositions correspond to the bounded faces of this polyhedron and minimal solutions must be vertices. We then identify cases with a unique minimal decomposition, and illustrate how our insights have consequences in the theory of submodular functions. Finally, we improve upon previous constructions of neural networks for a given convex CPWL function and apply our framework to obtain results in the nonconvex case.

☆ Low-Rank Continual Personalization of Diffusion Models

Łukasz Staniszewski, Katarzyna Zaleska, Kamil Deja

Recent personalization methods for diffusion models, such as Dreambooth, allow fine-tuning pre-trained models to generate new concepts. However, applying these techniques across multiple tasks in order to include, e.g., several new objects or styles, leads to mutual interference between their adapters. While recent studies attempt to mitigate this issue by combining trained adapters across tasks after fine-tuning, we adopt a more rigorous regime and investigate the personalization of large diffusion models under a continual learning scenario, where such interference leads to catastrophic forgetting of previous knowledge. To that end, we evaluate the na\"ive continual fine-tuning of customized models and compare this approach with three methods for consecutive adapters' training: sequentially merging new adapters, merging orthogonally initialized adapters, and updating only relevant parameters according to the task. In our experiments, we show that the proposed approaches mitigate forgetting when compared to the na\"ive approach.

☆ Wide Neural Networks Trained with Weight Decay Provably Exhibit Neural Collapse

Arthur Jacot, Peter Súkeník, Zihan Wang, Marco Mondelli

Deep neural networks (DNNs) at convergence consistently represent the training data in the last layer via a highly symmetric geometric structure referred to as neural collapse. This empirical evidence has spurred a line of theoretical research aimed at proving the emergence of neural collapse, mostly focusing on the unconstrained features model. Here, the features of the penultimate layer are free variables, which makes the model data-agnostic and, hence, puts into question its ability to capture DNN training. Our work addresses the issue, moving away from unconstrained features and studying DNNs that end with at least two linear layers. We first prove generic guarantees on neural collapse that assume (i) low training error and balancedness of the linear layers (for within-class variability collapse), and (ii) bounded conditioning of the features before the linear part (for orthogonality of class-means, as well as their alignment with weight matrices). We then show that such assumptions hold for gradient descent training with weight decay: (i) for networks with a wide first layer, we prove low training error and balancedness, and (ii) for solutions that are either nearly optimal or stable under large learning rates, we additionally prove the bounded conditioning. Taken together, our results are the first to show neural collapse in the end-to-end training of DNNs.

comment: 29 pages, 5 figures

☆ Improving the Sampling Strategy in KernelSHAP

Lars Henry Berge Olsen, Martin Jullum

Shapley values are a popular model-agnostic explanation framework for explaining predictions made by complex machine learning models. The framework provides feature contribution scores that sum to the predicted response and represent each feature's importance. The computation of exact Shapley values is computationally expensive due to estimating an exponential amount of non-trivial conditional expectations. The KernelSHAP framework enables us to approximate the Shapley values using a sampled subset of weighted conditional expectations. We propose three main novel contributions: a stabilizing technique to reduce the variance of the weights in the current state-of-the-art strategy, a novel weighing scheme that corrects the Shapley kernel weights based on sampled subsets, and a straightforward strategy that includes the important subsets and integrates them with the corrected Shapley kernel weights. We compare these new approximation strategies against existing ones by evaluating their Shapley value accuracy as a function of the number of subsets. The results demonstrate that our sampling strategies significantly enhance the accuracy of the approximated Shapley value explanations, making them more reliable in practical applications. This work provides valuable insights and practical recommendations for researchers and practitioners seeking to implement Shapley value-based explainability of their models.

☆ On the Optimization and Generalization of Two-layer Transformers with Sign Gradient Descent

Bingrui Li, Wei Huang, Andi Han, Zhanpeng Zhou, Taiji Suzuki, Jun Zhu, Jianfei Chen

The Adam optimizer is widely used for transformer optimization in practice, which makes understanding the underlying optimization mechanisms an important problem. However, due to the Adam's complexity, theoretical analysis of how it optimizes transformers remains a challenging task. Fortunately, Sign Gradient Descent (SignGD) serves as an effective surrogate for Adam. Despite its simplicity, theoretical understanding of how SignGD optimizes transformers still lags behind. In this work, we study how SignGD optimizes a two-layer transformer -- consisting of a softmax attention layer with trainable query-key parameterization followed by a linear layer -- on a linearly separable noisy dataset. We identify four stages in the training dynamics, each exhibiting intriguing behaviors. Based on the training dynamics, we prove the fast convergence but poor generalization of the learned transformer on the noisy dataset. We also show that Adam behaves similarly to SignGD in terms of both optimization and generalization in this setting. Additionally, we find that the poor generalization of SignGD is not solely due to data noise, suggesting that both SignGD and Adam requires high-quality data for real-world tasks. Finally, experiments on synthetic and real-world datasets empirically support our theoretical results.

comment: preprint

☆ Mastering Chinese Chess AI (Xiangqi) Without Search

Yu Chen, Juntong Lin, Zhichao Shu

We have developed a high-performance Chinese Chess AI that operates without reliance on search algorithms. This AI has demonstrated the capability to compete at a level commensurate with the top 0.1\% of human players. By eliminating the search process typically associated with such systems, this AI achieves a Queries Per Second (QPS) rate that exceeds those of systems based on the Monte Carlo Tree Search (MCTS) algorithm by over a thousandfold and surpasses those based on the AlphaBeta pruning algorithm by more than a hundredfold. The AI training system consists of two parts: supervised learning and reinforcement learning. Supervised learning provides an initial human-like Chinese chess AI, while reinforcement learning, based on supervised learning, elevates the strength of the entire AI to a new level. Based on this training system, we carried out enough ablation experiments and discovered that 1. The same parameter amount of Transformer architecture has a higher performance than CNN on Chinese chess; 2. Possible moves of both sides as features can greatly improve the training process; 3. Selective opponent pool, compared to pure self-play training, results in a faster improvement curve and a higher strength limit. 4. Value Estimation with Cutoff(VECT) improves the original PPO algorithm training process and we will give the explanation.

☆ Unsupervised Skill Discovery for Robotic Manipulation through Automatic Task Generation

Paul Jansonnie, Bingbing Wu, Julien Perez, Jan Peters

Learning skills that interact with objects is of major importance for robotic manipulation. These skills can indeed serve as an efficient prior for solving various manipulation tasks. We propose a novel Skill Learning approach that discovers composable behaviors by solving a large and diverse number of autonomously generated tasks. Our method learns skills allowing the robot to consistently and robustly interact with objects in its environment. The discovered behaviors are embedded in primitives which can be composed with Hierarchical Reinforcement Learning to solve unseen manipulation tasks. In particular, we leverage Asymmetric Self-Play to discover behaviors and Multiplicative Compositional Policies to embed them. We compare our method to Skill Learning baselines and find that our skills are more interactive. Furthermore, the learned skills can be used to solve a set of unseen manipulation tasks, in simulation as well as on a real robotic platform.

comment: Accepted at the 2024 IEEE-RAS International Conference on Humanoid Robots

☆ TimeCNN: Refining Cross-Variable Interaction on Time Point for Time Series Forecasting

Ao Hu, Dongkai Wang, Yong Dai, Shiyi Qi, Liangjian Wen, Jun Wang, Zhi Chen, Xun Zhou, Zenglin Xu, Jiang Duan

Time series forecasting is extensively applied across diverse domains. Transformer-based models demonstrate significant potential in modeling cross-time and cross-variable interaction. However, we notice that the cross-variable correlation of multivariate time series demonstrates multifaceted (positive and negative correlations) and dynamic progression over time, which is not well captured by existing Transformer-based models. To address this issue, we propose a TimeCNN model to refine cross-variable interactions to enhance time series forecasting. Its key innovation is timepoint-independent, where each time point has an independent convolution kernel, allowing each time point to have its independent model to capture relationships among variables. This approach effectively handles both positive and negative correlations and adapts to the evolving nature of variable relationships over time. Extensive experiments conducted on 12 real-world datasets demonstrate that TimeCNN consistently outperforms state-of-the-art models. Notably, our model achieves significant reductions in computational requirements (approximately 60.46%) and parameter count (about 57.50%), while delivering inference speeds 3 to 4 times faster than the benchmark iTransformer model

☆ Strong Model Collapse

Elvis Dohmatob, Yunzhen Feng, Julia Kempe

☆ Multimodal Fusion Strategies for Mapping Biophysical Landscape Features ECCV 2024

Lucia Gordon, Nico Lang, Catherine Ressijac, Andrew Davies

comment: 9 pages, 4 figures, ECCV 2024 Workshop in CV for Ecology

☆ Taming Gradient Oversmoothing and Expansion in Graph Neural Networks

MoonJeong Park, Dongwoo Kim

Oversmoothing has been claimed as a primary bottleneck for multi-layered graph neural networks (GNNs). Multiple analyses have examined how and why oversmoothing occurs. However, none of the prior work addressed how optimization is performed under the oversmoothing regime. In this work, we show the presence of $\textit{gradient oversmoothing}$ preventing optimization during training. We further analyze that GNNs with residual connections, a well-known solution to help gradient flow in deep architecture, introduce $\textit{gradient expansion}$, a phenomenon of the gradient explosion in diverse directions. Therefore, adding residual connections cannot be a solution for making a GNN deep. Our analysis reveals that constraining the Lipschitz bound of each layer can neutralize the gradient expansion. To this end, we provide a simple yet effective normalization method to prevent the gradient expansion. An empirical study shows that the residual GNNs with hundreds of layers can be efficiently trained with the proposed normalization without compromising performance. Additional studies show that the empirical observations corroborate our theoretical analysis.

☆ Physics-Informed GNN for non-linear constrained optimization: PINCO a solver for the AC-optimal power flow

Anna Varbella, Damien Briens, Blazhe Gjorgiev, Giuseppe Alessio D'Inverno, Giovanni Sansavini

The energy transition is driving the integration of large shares of intermittent power sources in the electric power grid. Therefore, addressing the AC optimal power flow (AC-OPF) effectively becomes increasingly essential. The AC-OPF, which is a fundamental optimization problem in power systems, must be solved more frequently to ensure the safe and cost-effective operation of power systems. Due to its non-linear nature, AC-OPF is often solved in its linearized form, despite inherent inaccuracies. Non-linear solvers, such as the interior point method, are typically employed to solve the full OPF problem. However, these iterative methods may not converge for large systems and do not guarantee global optimality. This work explores a physics-informed graph neural network, PINCO, to solve the AC-OPF. We demonstrate that this method provides accurate solutions in a fraction of the computational time when compared to the established non-linear programming solvers. Remarkably, PINCO generalizes effectively across a diverse set of loading conditions in the power system. We show that our method can solve the AC-OPF without violating inequality constraints. Furthermore, it can function both as a solver and as a hybrid universal function approximator. Moreover, the approach can be easily adapted to different power systems with minimal adjustments to the hyperparameters, including systems with multiple generators at each bus. Overall, this work demonstrates an advancement in the field of power system optimization to tackle the challenges of the energy transition. The code and data utilized in this paper are available at https://anonymous.4open.science/r/opf_pinn_iclr-B83E/.

☆ Learning Interpretable Hierarchical Dynamical Systems Models from Time Series Data

Manuel Brenner, Elias Weber, Georgia Koppe, Daniel Durstewitz

In science, we are often interested in obtaining a generative model of the underlying system dynamics from observed time series. While powerful methods for dynamical systems reconstruction (DSR) exist when data come from a single domain, how to best integrate data from multiple dynamical regimes and leverage it for generalization is still an open question. This becomes particularly important when individual time series are short, and group-level information may help to fill in for gaps in single-domain data. At the same time, averaging is not an option in DSR, as it will wipe out crucial dynamical properties (e.g., limit cycles in one domain vs. chaos in another). Hence, a framework is needed that enables to efficiently harvest group-level (multi-domain) information while retaining all single-domain dynamical characteristics. Here we provide such a hierarchical approach and showcase it on popular DSR benchmarks, as well as on neuroscientific and medical time series. In addition to faithful reconstruction of all individual dynamical regimes, our unsupervised methodology discovers common low-dimensional feature spaces in which datasets with similar dynamics cluster. The features spanning these spaces were further dynamically highly interpretable, surprisingly in often linear relation to control parameters that govern the dynamics of the underlying system. Finally, we illustrate transfer learning and generalization to new parameter regimes.

comment: Preprint

☆ FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

Haokun Chen, Hang Li, Yao Zhang, Gengyuan Zhang, Jinhe Bi, Philip Torr, Jindong Gu, Denis Krompass, Volker Tresp

♻ ☆ mDPO: Conditional Preference Optimization for Multimodal Large Language Models EMNLP 2024

Fei Wang, Wenxuan Zhou, James Y. Huang, Nan Xu, Sheng Zhang, Hoifung Poon, Muhao Chen

comment: Accepted to EMNLP 2024 Main Conference. Project website: https://feiwang96.github.io/mDPO

♻ ☆ SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)

Matthieu Meeus, Igor Shilov, Shubham Jain, Manuel Faysse, Marek Rei, Yves-Alexandre de Montjoye

♻ ☆ Generative Parameter-Efficient Fine-Tuning

Chinmay Savadikar, Xi Song, Tianfu Wu

comment: Project page and code: https://savadikarc.github.io/gift

♻ ☆ Jogging the Memory of Unlearned LLMs Through Targeted Relearning Attack

Shengyuan Hu, Yiwei Fu, Zhiwei Steven Wu, Virginia Smith

comment: 26 pages, 5 figures, 7 tables

♻ ☆ Learning Successor Features with Distributed Hebbian Temporal Memory

Evgenii Dzhivelikian, Petr Kuderov, Aleksandr I. Panov

This paper presents a novel approach to address the challenge of online temporal memory learning for decision-making under uncertainty in non-stationary, partially observable environments. The proposed algorithm, Distributed Hebbian Temporal Memory (DHTM), is based on factor graph formalism and a multicomponent neuron model. DHTM aims to capture sequential data relationships and make cumulative predictions about future observations, forming Successor Features (SF). Inspired by neurophysiological models of the neocortex, the algorithm utilizes distributed representations, sparse transition matrices, and local Hebbian-like learning rules to overcome the instability and slow learning process of traditional temporal memory algorithms like RNN and HMM. Experimental results demonstrate that DHTM outperforms LSTM and a biologically inspired HMM-like algorithm, CSCG, in the case of non-stationary datasets. Our findings suggest that DHTM is a promising approach for addressing the challenges of online sequence learning and planning in dynamic environments.

comment: 20 pages, 7 figures

♻ ☆ Full Line Code Completion: Bringing AI to Desktop

Anton Semenkin, Vitaliy Bibaev, Yaroslav Sokolov, Kirill Krylov, Alexey Kalina, Anna Khannanova, Danila Savenkov, Darya Rovdo, Igor Davidenko, Kirill Karnaukhov, Maxim Vakhrushev, Mikhail Kostyukov, Mikhail Podvitskii, Petr Surkov, Yaroslav Golubev, Nikita Povarov, Timofey Bryksin

In recent years, several industrial solutions for the problem of multi-token code completion appeared, each making a great advance in the area but mostly focusing on cloud-based runtime and avoiding working on the end user's device. In this work, we describe our approach for building a multi-token code completion feature for the JetBrains' IntelliJ Platform, which we call Full Line Code Completion. The feature suggests only syntactically correct code and works fully locally, i.e., data querying and the generation of suggestions happens on the end user's machine. We share important time and memory-consumption restrictions, as well as design principles that a code completion engine should satisfy. Working entirely on the end user's device, our code completion engine enriches user experience while being not only fast and compact but also secure. We share a number of useful techniques to meet the stated development constraints and also describe offline and online evaluation pipelines that allowed us to make better decisions. Our online evaluation shows that the usage of the tool leads to 1.3 times more Python code in the IDE being produced by code completion. The described solution was initially started with a help of researchers and was then bundled into all JetBrains IDEs where it is now used by millions of users. Thus, we believe that this work is useful for bridging academia and industry, providing researchers with the knowledge of what happens when complex research-based solutions are integrated into real products.

comment: 12 pages, 4 figures

♻ ☆ Stateful Large Language Model Serving with Pensieve

Lingfan Yu, Jinkun Lin, Jinyang Li

♻ ☆ Online Convex Optimization with a Separation Oracle

Zakaria Mhammedi

In this paper, we introduce a new projection-free algorithm for Online Convex Optimization (OCO) with a state-of-the-art regret guarantee among separation-based algorithms. Existing projection-free methods based on the classical Frank-Wolfe algorithm achieve a suboptimal regret bound of $O(T^{3/4})$, while more recent separation-based approaches guarantee a regret bound of $O(\kappa \sqrt{T})$, where $\kappa$ denotes the asphericity of the feasible set, defined as the ratio of the radii of the containing and contained balls. However, for ill-conditioned sets, $\kappa$ can be arbitrarily large, potentially leading to poor performance. Our algorithm achieves a regret bound of $\widetilde{O}(\sqrt{dT} + \kappa d)$, while requiring only $\widetilde{O}(1)$ calls to a separation oracle per round. Crucially, the main term in the bound, $\widetilde{O}(\sqrt{d T})$, is independent of $\kappa$, addressing the limitations of previous methods. Additionally, as a by-product of our analysis, we recover the $O(\kappa \sqrt{T})$ regret bound of existing OCO algorithms with a more straightforward analysis and improve the regret bound for projection-free online exp-concave optimization. Finally, for constrained stochastic convex optimization, we achieve a state-of-the-art convergence rate of $\widetilde{O}(\sigma/\sqrt{T} + \kappa d/T)$, where $\sigma$ represents the noise in the stochastic gradients, while requiring only $\widetilde{O}(1)$ calls to a separation oracle per iteration.

♻ ☆ CAnDOIT: Causal Discovery with Observational and Interventional Data from Time-Series

Luca Castri, Sariah Mghames, Marc Hanheide, Nicola Bellotto

The study of cause-and-effect is of the utmost importance in many branches of science, but also for many practical applications of intelligent systems. In particular, identifying causal relationships in situations that include hidden factors is a major challenge for methods that rely solely on observational data for building causal models. This paper proposes CAnDOIT, a causal discovery method to reconstruct causal models using both observational and interventional time-series data. The use of interventional data in the causal analysis is crucial for real-world applications, such as robotics, where the scenario is highly complex and observational data alone are often insufficient to uncover the correct causal structure. Validation of the method is performed initially on randomly generated synthetic models and subsequently on a well-known benchmark for causal structure learning in a robotic manipulation environment. The experiments demonstrate that the approach can effectively handle data from interventions and exploit them to enhance the accuracy of the causal analysis. A Python implementation of CAnDOIT has also been developed and is publicly available on GitHub: https://github.com/lcastri/causalflow.

comment: Published in Advanced Intelligent Systems

♻ ☆ Optimal Aggregation of Prediction Intervals under Unsupervised Domain Shift

Jiawei Ge, Debarghya Mukherjee, Jianqing Fan

As machine learning models are increasingly deployed in dynamic environments, it becomes paramount to assess and quantify uncertainties associated with distribution shifts. A distribution shift occurs when the underlying data-generating process changes, leading to a deviation in the model's performance. The prediction interval, which captures the range of likely outcomes for a given prediction, serves as a crucial tool for characterizing uncertainties induced by their underlying distribution. In this paper, we propose methodologies for aggregating prediction intervals to obtain one with minimal width and adequate coverage on the target domain under unsupervised domain shift, under which we have labeled samples from a related source domain and unlabeled covariates from the target domain. Our analysis encompasses scenarios where the source and the target domain are related via i) a bounded density ratio, and ii) a measure-preserving transformation. Our proposed methodologies are computationally efficient and easy to implement. Beyond illustrating the performance of our method through real-world datasets, we also delve into the theoretical details. This includes establishing rigorous theoretical guarantees, coupled with finite sample bounds, regarding the coverage and width of our prediction intervals. Our approach excels in practical applications and is underpinned by a solid theoretical framework, ensuring its reliability and effectiveness across diverse contexts.

♻ ☆ Principal-Agent Reinforcement Learning: Orchestrating AI Agents with Contracts

Dima Ivanov, Paul Dütting, Inbal Talgam-Cohen, Tonghan Wang, David C. Parkes

The increasing deployment of AI is shaping the future landscape of the internet, which is set to become an integrated ecosystem of AI agents. Orchestrating the interaction among AI agents necessitates decentralized, self-sustaining mechanisms that harmonize the tension between individual interests and social welfare. In this paper we tackle this challenge by synergizing reinforcement learning with principal-agent theory from economics. Taken separately, the former allows unrealistic freedom of intervention, while the latter struggles to scale in sequential settings. Combining them achieves the best of both worlds. We propose a framework where a principal guides an agent in a Markov Decision Process (MDP) using a series of contracts, which specify payments by the principal based on observable outcomes of the agent's actions. We present and analyze a meta-algorithm that iteratively optimizes the policies of the principal and agent, showing its equivalence to a contraction operator on the principal's Q-function, and its convergence to subgame-perfect equilibrium. We then scale our algorithm with deep Q-learning and analyze its convergence in the presence of approximation error, both theoretically and through experiments with randomly generated binary game-trees. Extending our framework to multiple agents, we apply our methodology to the combinatorial Coin Game. Addressing this multi-agent sequential social dilemma is a promising first step toward scaling our approach to more complex, real-world instances.

♻ ☆ Creative Beam Search: LLM-as-a-Judge For Improving Response Generation

Giorgio Franceschelli, Mirco Musolesi

comment: Presented as a short paper at the 15th International Conference on Computational Creativity (ICCC'24)

♻ ☆ Forest Proximities for Time Series

Ben Shaw, Jake Rhodes, Soukaina Filali Boubrahimi, Kevin R. Moon

RF-GAP has recently been introduced as an improved random forest proximity measure. In this paper, we present PF-GAP, an extension of RF-GAP proximities to proximity forests, an accurate and efficient time series classification model. We use the forest proximities in connection with Multi-Dimensional Scaling to obtain vector embeddings of univariate time series, comparing the embeddings to those obtained using various time series distance measures. We also use the forest proximities alongside Local Outlier Factors to investigate the connection between misclassified points and outliers, comparing with nearest neighbor classifiers which use time series distance measures. We show that the forest proximities may exhibit a stronger connection between misclassified points and outliers than nearest neighbor classifiers.

♻ ☆ Training Foundation Models as Data Compression: On Information, Model Weights and Copyright Law

Giorgio Franceschelli, Claudia Cevenini, Mirco Musolesi

The training process of foundation models as for other classes of deep learning systems is based on minimizing the reconstruction error over a training set. For this reason, they are susceptible to the memorization and subsequent reproduction of training samples. In this paper, we introduce a training-as-compressing perspective, wherein the model's weights embody a compressed representation of the training data. From a copyright standpoint, this point of view implies that the weights could be considered a reproduction or a derivative work of a potentially protected set of works. We investigate the technical and legal challenges that emerge from this framing of the copyright of outputs generated by foundation models, including their implications for practitioners and researchers. We demonstrate that adopting an information-centric approach to the problem presents a promising pathway for tackling these emerging complex legal issues.

comment: Spotlight presentation at GenLaw'24, see https://www.genlaw.org/2024-icml-papers#training-foundation-models-as-data-compression-on-information-model-weights-and-copyright-law

♻ ☆ MetaMetrics: Calibrating Metrics For Generation Tasks Using Human Preferences

Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, Derry Tanti Wijaya

comment: Preprint

♻ ☆ Machine Learning Based Optimal Design of Fibrillar Adhesives

Mohammad Shojaeifard, Matteo Ferraresso, Alessandro Lucantonio, Mattia Bacca

Fibrillar adhesion, observed in animals like beetles, spiders, and geckos, relies on nanoscopic or microscopic fibrils to enhance surface adhesion via 'contact splitting.' This concept has inspired engineering applications across robotics, transportation, and medicine. Recent studies suggest that functional grading of fibril properties can improve adhesion, but this is a complex design challenge that has only been explored in simplified geometries. While machine learning (ML) has gained traction in adhesive design, no previous attempts have targeted fibril-array scale optimization. In this study, we propose an ML-based tool that optimizes the distribution of fibril compliance to maximize adhesive strength. Our tool, featuring two deep neural networks (DNNs), recovers previous design results for simple geometries and introduces novel solutions for complex configurations. The Predictor DNN estimates adhesive strength based on random compliance distributions, while the Designer DNN optimizes compliance for maximum strength using gradient-based optimization. Our method significantly reduces test error and accelerates the optimization process, offering a high-performance solution for designing fibrillar adhesives and micro-architected materials aimed at fracture resistance by achieving equal load sharing (ELS).

♻ ☆ Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks

Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion

We show that even the most recent safety-aligned LLMs are not robust to simple adaptive jailbreaking attacks. First, we demonstrate how to successfully leverage access to logprobs for jailbreaking: we initially design an adversarial prompt template (sometimes adapted to the target LLM), and then we apply random search on a suffix to maximize a target logprob (e.g., of the token "Sure"), potentially with multiple restarts. In this way, we achieve 100% attack success rate -- according to GPT-4 as a judge -- on Vicuna-13B, Mistral-7B, Phi-3-Mini, Nemotron-4-340B, Llama-2-Chat-7B/13B/70B, Llama-3-Instruct-8B, Gemma-7B, GPT-3.5, GPT-4o, and R2D2 from HarmBench that was adversarially trained against the GCG attack. We also show how to jailbreak all Claude models -- that do not expose logprobs -- via either a transfer or prefilling attack with a 100% success rate. In addition, we show how to use random search on a restricted set of tokens for finding trojan strings in poisoned models -- a task that shares many similarities with jailbreaking -- which is the algorithm that brought us the first place in the SaTML'24 Trojan Detection Competition. The common theme behind these attacks is that adaptivity is crucial: different models are vulnerable to different prompting templates (e.g., R2D2 is very sensitive to in-context learning prompts), some models have unique vulnerabilities based on their APIs (e.g., prefilling for Claude), and in some settings, it is crucial to restrict the token search space based on prior knowledge (e.g., for trojan detection). For reproducibility purposes, we provide the code, logs, and jailbreak artifacts in the JailbreakBench format at https://github.com/tml-epfl/llm-adaptive-attacks.

comment: Updates in the v3: GPT-4o and Claude 3.5 Sonnet results, improved writing. Updates in the v2: more models (Llama3, Phi-3, Nemotron-4-340B), jailbreak artifacts for all attacks are available, evaluation with different judges (Llama-3-70B and Llama Guard 2), more experiments (convergence plots over iterations, ablation on the suffix length for random search), examples of jailbroken generation

♻ ☆ Efficient Model-Agnostic Multi-Group Equivariant Networks

Razan Baltaji, Sourya Basu, Lav R. Varshney

♻ ☆ When "A Helpful Assistant" Is Not Really Helpful: Personas in System Prompts Do Not Improve Performances of Large Language Models

Mingqian Zheng, Jiaxin Pei, Lajanugen Logeswaran, Moontae Lee, David Jurgens

♻ ☆ Learning to Steer Markovian Agents under Model Uncertainty

Jiawei Huang, Vinzenz Thoma, Zebang Shen, Heinrich H. Nax, Niao He

Designing incentives for an adapting population is a ubiquitous problem in a wide array of economic applications and beyond. In this work, we study how to design additional rewards to steer multi-agent systems towards desired policies \emph{without} prior knowledge of the agents' underlying learning dynamics. Motivated by the limitation of existing works, we consider a new and general category of learning dynamics called \emph{Markovian agents}. We introduce a model-based non-episodic Reinforcement Learning (RL) formulation for our steering problem. Importantly, we focus on learning a \emph{history-dependent} steering strategy to handle the inherent model uncertainty about the agents' learning dynamics. We introduce a novel objective function to encode the desiderata of achieving a good steering outcome with reasonable cost. Theoretically, we identify conditions for the existence of steering strategies to guide agents to the desired policies. Complementing our theoretical contributions, we provide empirical algorithms to approximately solve our objective, which effectively tackles the challenge in learning history-dependent strategies. We demonstrate the efficacy of our algorithms through empirical evaluations.

comment: 34 Pages

♻ ☆ Deep Fusion: Capturing Dependencies in Contrastive Learning via Transformer Projection Heads

Huanran Li, Daniel Pimentel-Alarcón

Contrastive Learning (CL) has emerged as a powerful method for training feature extraction models using unlabeled data. Recent studies suggest that incorporating a linear projection head post-backbone significantly enhances model performance. In this work, we investigate the use of a transformer model as a projection head within the CL framework, aiming to exploit the transformer's capacity for capturing long-range dependencies across embeddings to further improve performance. Our key contributions are fourfold: First, we introduce a novel application of transformers in the projection head role for contrastive learning, marking the first endeavor of its kind. Second, our experiments reveal a compelling "Deep Fusion" phenomenon where the attention mechanism progressively captures the correct relational dependencies among samples from the same class in deeper layers. Third, we provide a theoretical framework that explains and supports this "Deep Fusion" behavior. Finally, we demonstrate through experimental results that our model achieves superior performance compared to the existing approach of using a feed-forward layer.

comment: 10 pages, 2 figures

♻ ☆ Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding

Zachary Ankner, Rishab Parthasarathy, Aniruddha Nrusimha, Christopher Rinard, Jonathan Ragan-Kelley, William Brandon

To combat the memory bandwidth-bound nature of autoregressive LLM inference, previous research has proposed the speculative decoding frame-work. To perform speculative decoding, a small draft model proposes candidate continuations of the input sequence that are then verified in parallel by the base model. One way to specify the draft model, as used in the recent Medusa decoding framework, is as a collection of lightweight heads, called draft heads, that operate on the base model's hidden states. To date, all existing draft heads have been sequentially independent, meaning that they speculate tokens in the candidate continuation independently of any preceding tokens in the candidate continuation. In this work, we propose Hydra heads: a sequentially-dependent drop-in replacement for standard draft heads that significantly improves the accuracy of draft head speculation. We further explore the design space of Hydra head training objectives and architectures, and propose a carefully tuned Hydra head recipe, which we call Hydra++, that improves decoding throughput by up to 1.31x and 2.70x compared to Medusa decoding and autoregressive de-coding respectively. Overall, Hydra heads are a simple and well-motivated intervention on standard draft heads that significantly improve the end-to-end speed of draft head-based speculative decoding. We make our code publicly available at https://github.com/zankner/Hydra.

♻ ☆ The SkipSponge Attack: Sponge Weight Poisoning of Deep Neural Networks

Jona te Lintelo, Stefanos Koffas, Stjepan Picek

Sponge attacks aim to increase the energy consumption and computation time of neural networks. In this work, we present a novel sponge attack called SkipSponge. SkipSponge is the first sponge attack that is performed directly on the parameters of a pre-trained model using only a few data samples. Our experiments show that SkipSponge can successfully increase the energy consumption of image classification models, GANs, and autoencoders requiring fewer samples than the state-of-the-art (Sponge Poisoning). We show that poisoning defenses are ineffective if not adjusted specifically for the defense against SkipSponge (i.e., they decrease target layer bias values). Our work shows that SkipSponge is more effective on the GANs and the autoencoders than Sponge Poisoning. Additionally, SkipSponge is stealthier than Sponge Poisoning as it does not require significant changes in the victim model's weights. Our experiments indicate that SkipSponge can be performed even when an attacker has access to only 1% of the entire dataset and reaches up to 13% energy increase.

♻ ☆ Robust Multimodal Learning with Missing Modalities via Parameter-Efficient Adaptation

Md Kaykobad Reza, Ashley Prater-Bennette, M. Salman Asif

comment: Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 28 pages, 6 figures, 17 tables

♻ ☆ Preventing Collapse in Contrastive Learning with Orthonormal Prototypes (CLOP)

Huanran Li, Manh Nguyen, Daniel Pimentel-Alarcón

Contrastive learning has emerged as a powerful method in deep learning, excelling at learning effective representations through contrasting samples from different distributions. However, neural collapse, where embeddings converge into a lower-dimensional space, poses a significant challenge, especially in semi-supervised and self-supervised setups. In this paper, we first theoretically analyze the effect of large learning rates on contrastive losses that solely rely on the cosine similarity metric, and derive a theoretical bound to mitigate this collapse. {Building on these insights, we propose CLOP, a novel semi-supervised loss function designed to prevent neural collapse by promoting the formation of orthogonal linear subspaces among class embeddings.} Unlike prior approaches that enforce a simplex ETF structure, CLOP focuses on subspace separation, leading to more distinguishable embeddings. Through extensive experiments on real and synthetic datasets, we demonstrate that CLOP enhances performance, providing greater stability across different learning rates and batch sizes.

comment: 17 pages, 8 figures

♻ ☆ NoSENSE: Learned unrolled cardiac MRI reconstruction without explicit sensitivity maps MICCAI

Felix Frederik Zimmermann, Andreas Kofler

comment: Accepted at MICCAI STACOM 2023

♻ ☆ Representation noising effectively prevents harmful fine-tuning on LLMs

Domenic Rosati, Jan Wehner, Kai Williams, Łukasz Bartoszcze, David Atanasov, Robie Gonzales, Subhabrata Majumdar, Carsten Maple, Hassan Sajjad, Frank Rudzicz

comment: Published in NeurIPs 2024

♻ ☆ Differentiable and Learnable Wireless Simulation with Geometric Transformers

Thomas Hehn, Markus Peschl, Tribhuvanesh Orekondy, Arash Behboodi, Johann Brehmer

Modelling the propagation of electromagnetic wireless signals is critical for designing modern communication systems. Wireless ray tracing simulators model signal propagation based on the 3D geometry and other scene parameters, but their accuracy is fundamentally limited by underlying modelling assumptions and correctness of parameters. In this work, we introduce Wi-GATr, a fully-learnable neural simulation surrogate designed to predict the channel observations based on scene primitives (e.g., surface mesh, antenna position and orientation). Recognizing the inherently geometric nature of these primitives, Wi-GATr leverages an equivariant Geometric Algebra Transformer that operates on a tokenizer specifically tailored for wireless simulation. We evaluate our approach on a range of tasks (i.e., signal strength and delay spread prediction, receiver localization, and geometry reconstruction) and find that Wi-GATr is accurate, fast, sample-efficient, and robust to symmetry-induced transformations. Remarkably, we find our results also translate well to the real world: Wi-GATr demonstrates more than 35% lower error than hybrid techniques, and 70% lower error than a calibrated wireless tracer.

♻ ☆ Efficient Gradient Estimation of Variational Quantum Circuits with Lie Algebraic Symmetries

Mohsen Heidari, Masih Mozakka, Wojciech Szpankowski

Hybrid quantum-classical optimization and learning strategies are among the most promising approaches to harnessing quantum information or gaining a quantum advantage over classical methods. However, efficient estimation of the gradient of the objective function in such models remains a challenge due to several factors including the exponential dimensionality of the Hilbert spaces, and information loss of quantum measurements. In this work, we developed an efficient framework that makes the Hadamard test efficiently applicable to gradient estimation for a broad range of quantum systems, an advance that had been wanting from the outset. Under certain mild structural assumptions, the gradient is estimated with the measurement shots that scale logarithmically with the number of parameters and with polynomial classical and quantum time. This is an exponential reduction in the measurement cost and polynomial speed up in time compared to existing works. The structural assumptions are (1) the dimension of the dynamical Lie algebra is polynomial in the number of qubits, and (2) the observable has a bounded Hilbert-Schmidt norm.

comment: 39 pages

♻ ☆ Fine-Tuning and Prompt Optimization: Two Great Steps that Work Better Together EMNLP 2024

Dilara Soylu, Christopher Potts, Omar Khattab

comment: EMNLP 2024

♻ ☆ Dr. Jekyll and Mr. Hyde: Two Faces of LLMs

Matteo Gioele Collu, Tom Janssen-Groesbeek, Stefanos Koffas, Mauro Conti, Stjepan Picek

Recently, we have witnessed a rise in the use of Large Language Models (LLMs), especially in applications like chatbots. Safety mechanisms are implemented to prevent improper responses from these chatbots. In this work, we bypass these measures for ChatGPT and Gemini by making them impersonate complex personas with personality characteristics that are not aligned with a truthful assistant. First, we create elaborate biographies of these personas, which we then use in a new session with the same chatbots. Our conversations then follow a role-play style to elicit prohibited responses. Using personas, we show that prohibited responses are provided, making it possible to obtain unauthorized, illegal, or harmful information in both ChatGPT and Gemini. We also introduce several ways of activating such adversarial personas, showing that both chatbots are vulnerable to this attack. With the same principle, we introduce two defenses that push the model to interpret trustworthy personalities and make it more robust against such attacks.

♻ ☆ A Neural-Evolutionary Algorithm for Autonomous Transit Network Design

Andrew Holliday, Gregory Dudek

Planning a public transit network is a challenging optimization problem, but essential in order to realize the benefits of autonomous buses. We propose a novel algorithm for planning networks of routes for autonomous buses. We first train a graph neural net model as a policy for constructing route networks, and then use the policy as one of several mutation operators in a evolutionary algorithm. We evaluate this algorithm on a standard set of benchmarks for transit network design, and find that it outperforms the learned policy alone by up to 20% and a plain evolutionary algorithm approach by up to 53% on realistic benchmark instances.

comment: Copyright 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. arXiv admin note: text overlap with arXiv:2306.00720

♻ ☆ Memory-Enhanced Neural Solvers for Efficient Adaptation in Combinatorial Optimization

Felix Chalumeau, Refiloe Shabe, Noah De Nicola, Arnu Pretorius, Thomas D. Barrett, Nathan Grinsztajn

Combinatorial Optimization is crucial to numerous real-world applications, yet still presents challenges due to its (NP-)hard nature. Amongst existing approaches, heuristics often offer the best trade-off between quality and scalability, making them suitable for industrial use. While Reinforcement Learning (RL) offers a flexible framework for designing heuristics, its adoption over handcrafted heuristics remains incomplete within industrial solvers. Existing learned methods still lack the ability to adapt to specific instances and fully leverage the available computational budget. The current best methods either rely on a collection of pre-trained policies, or on data-inefficient fine-tuning; hence failing to fully utilize newly available information within the constraints of the budget. In response, we present MEMENTO, an approach that leverages memory to improve the adaptation of neural solvers at inference time. MEMENTO enables updating the action distribution dynamically based on the outcome of previous decisions. We validate its effectiveness on benchmark problems, in particular Traveling Salesman and Capacitated Vehicle Routing, demonstrating its superiority over tree-search and policy-gradient fine-tuning; and showing it can be zero-shot combined with diversity-based solvers. We successfully train all RL auto-regressive solvers on large instances, and show that MEMENTO can scale and is data-efficient. Overall, MEMENTO enables to push the state-of-the-art on 11 out of 12 evaluated tasks.

♻ ☆ Investigating Guiding Information for Adaptive Collocation Point Sampling in PINNs CCS

Jose Florido, He Wang, Amirul Khan, Peter K. Jimack

Physics-informed neural networks (PINNs) provide a means of obtaining approximate solutions of partial differential equations and systems through the minimisation of an objective function which includes the evaluation of a residual function at a set of collocation points within the domain. The quality of a PINNs solution depends upon numerous parameters, including the number and distribution of these collocation points. In this paper we consider a number of strategies for selecting these points and investigate their impact on the overall accuracy of the method. In particular, we suggest that no single approach is likely to be "optimal" but we show how a number of important metrics can have an impact in improving the quality of the results obtained when using a fixed number of residual evaluations. We illustrate these approaches through the use of two benchmark test problems: Burgers' equation and the Allen-Cahn equation.

comment: 15 pages, 8 figures, 2 tables. Published in the conference proceedings of the International Conference on Computational Science (ICCS) 2024. Replacement to correct a typo regarding the value of viscosity listed in the captions

♻ ☆ LayerKV: Optimizing Large Language Model Serving with Layer-wise KV Cache Management

Yi Xiong, Hao Wu, Changxu Shao, Ziqing Wang, Rui Zhang, Yuhong Guo, Junping Zhao, Ke Zhang, Zhenxuan Pan

comment: 11 pages, 7 figures, 1 table

♻ ☆ On the Convergence of Hermitian Dynamic Mode Decomposition

Nicolas Boullé, Matthew J. Colbrook

We study the convergence of Hermitian Dynamic Mode Decomposition (DMD) to the spectral properties of self-adjoint Koopman operators. Hermitian DMD is a data-driven method that approximates the Koopman operator associated with an unknown nonlinear dynamical system, using discrete-time snapshots. This approach preserves the self-adjointness of the operator in its finite-dimensional approximations. \rev{We prove that, under suitably broad conditions, the spectral measures corresponding to the eigenvalues and eigenfunctions computed by Hermitian DMD converge to those of the underlying Koopman operator}. This result also applies to skew-Hermitian systems (after multiplication by $i$), applicable to generators of continuous-time measure-preserving systems. Along the way, we establish a general theorem on the convergence of spectral measures for finite sections of self-adjoint operators, including those that are unbounded, which is of independent interest to the wider spectral community. We numerically demonstrate our results by applying them to two-dimensional Schr\"odinger equations.

comment: 24 pages, 4 figures. arXiv admin note: text overlap with arXiv:2312.00137

♻ ☆ Decoding Intelligence: A Framework for Certifying Knowledge Comprehension in LLMs

Isha Chaudhary, Vedaant V. Jain, Gagandeep Singh

♻ ☆ Transition Path Sampling with Improved Off-Policy Training of Diffusion Path Samplers

Kiyoung Seong, Seonghyun Park, Seonghwan Kim, Woo Youn Kim, Sungsoo Ahn

Understanding transition pathways between meta-stable states in molecular systems is crucial to advance material design and drug discovery. However, unbiased molecular dynamics simulations are computationally infeasible due to the high energy barriers separating these states. Although recent machine learning techniques offer potential solutions, they are often limited to simple systems or rely on collective variables (CVs) derived from costly domain expertise. In this paper, we introduce a novel approach that trains diffusion path samplers (DPS) for transition path sampling (TPS) without the need for CVs. We recast the problem as an amortized sampling of the target path measure, minimizing the log-variance divergence between the path measure induced by our DPS and the target path measure. To ensure scalability for high-dimensional tasks, we introduce (1) a new off-policy training objective based on learning control variates with replay buffers and (2) a scale-based equivariant parameterization of the bias forces. We evaluate our approach, coined TPS-DPS, on a synthetic double-well potential and three peptides: Alanine Dipeptide, Polyproline Helix, and Chignolin. Results show that our approach produces more realistic and diverse transition pathways compared to existing baselines.

comment: 10 pages, 8 figures, 1 tables

♻ ☆ WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models NeurIPS 2024

Peng Wang, Zexi Li, Ningyu Zhang, Ziwen Xu, Yunzhi Yao, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen

comment: NeurIPS 2024

♻ ☆ Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts

Junmo Kang, Leonid Karlinsky, Hongyin Luo, Zhen Wang, Jacob Hansen, James Glass, David Cox, Rameswar Panda, Rogerio Feris, Alan Ritter

♻ ☆ An Elementary Predictor Obtaining $2\sqrt{T}+1$ Distance to Calibration

Eshwar Ram Arunachaleswaran, Natalie Collina, Aaron Roth, Mirah Shi

Blasiok et al. [2023] proposed distance to calibration as a natural measure of calibration error that unlike expected calibration error (ECE) is continuous. Recently, Qiao and Zheng [2024] gave a non-constructive argument establishing the existence of an online predictor that can obtain $O(\sqrt{T})$ distance to calibration in the adversarial setting, which is known to be impossible for ECE. They leave as an open problem finding an explicit, efficient algorithm. We resolve this problem and give an extremely simple, efficient, deterministic algorithm that obtains distance to calibration error at most $2\sqrt{T}+1$.

♻ ☆ Function-Space MCMC for Bayesian Wide Neural Networks

Lucia Pezzetti, Stefano Favaro, Stefano Peluchetti

Bayesian Neural Networks represent a fascinating confluence of deep learning and probabilistic reasoning, offering a compelling framework for understanding uncertainty in complex predictive models. In this paper, we investigate the use of the preconditioned Crank-Nicolson algorithm and its Langevin version to sample from the reparametrised posterior distribution of the weights as the widths of Bayesian Neural Networks grow larger. In addition to being robust in the infinite-dimensional setting, we prove that the acceptance probabilities of the proposed methods approach 1 as the width of the network increases, independently of any stepsize tuning. Moreover, we examine and compare how the mixing speeds of the underdamped Langevin Monte Carlo, the preconditioned Crank-Nicolson and the preconditioned Crank-Nicolson Langevin samplers are influenced by changes in the network width in some real-world cases. Our findings suggest that, in wide Bayesian Neural Networks configurations, the preconditioned Crank-Nicolson method allows for more efficient sampling of the reparametrised posterior distribution, as evidenced by a higher effective sample size and improved diagnostic results compared with the other analysed algorithms.

♻ ☆ Data-Centric Foundation Models in Computational Healthcare: A Survey

Yunkun Zhang, Jin Gao, Zheling Tan, Lingfeng Zhou, Kexin Ding, Mu Zhou, Shaoting Zhang, Dequan Wang

The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. The interactive nature of these models, guided by pre-training data and human instructions, has ignited a data-centric AI paradigm that emphasizes better data characterization, quality, and scale. In healthcare AI, obtaining and processing high-quality clinical data records has been a longstanding challenge, ranging from data quantity, annotation, patient privacy, and ethics. In this survey, we investigate a wide range of data-centric approaches in the FM era (from model pre-training to inference) towards improving the healthcare workflow. We discuss key perspectives in AI security, assessment, and alignment with human values. Finally, we offer a promising outlook of FM-based analytics to enhance the performance of patient outcome and clinical workflow in the evolving landscape of healthcare and medicine. We provide an up-to-date list of healthcare-related foundation models and datasets at https://github.com/Yunkun-Zhang/Data-Centric-FM-Healthcare .

comment: Survey content updated to include recent research work and progress

♻ ☆ Geodesic Optimization for Predictive Shift Adaptation on EEG data

Apolline Mellot, Antoine Collas, Sylvain Chevallier, Alexandre Gramfort, Denis A. Engemann

Electroencephalography (EEG) data is often collected from diverse contexts involving different populations and EEG devices. This variability can induce distribution shifts in the data $X$ and in the biomedical variables of interest $y$, thus limiting the application of supervised machine learning (ML) algorithms. While domain adaptation (DA) methods have been developed to mitigate the impact of these shifts, such methods struggle when distribution shifts occur simultaneously in $X$ and $y$. As state-of-the-art ML models for EEG represent the data by spatial covariance matrices, which lie on the Riemannian manifold of Symmetric Positive Definite (SPD) matrices, it is appealing to study DA techniques operating on the SPD manifold. This paper proposes a novel method termed Geodesic Optimization for Predictive Shift Adaptation (GOPSA) to address test-time multi-source DA for situations in which source domains have distinct $y$ distributions. GOPSA exploits the geodesic structure of the Riemannian manifold to jointly learn a domain-specific re-centering operator representing site-specific intercepts and the regression model. We performed empirical benchmarks on the cross-site generalization of age-prediction models with resting-state EEG data from a large multi-national dataset (HarMNqEEG), which included $14$ recording sites and more than $1500$ human participants. Compared to state-of-the-art methods, our results showed that GOPSA achieved significantly higher performance on three regression metrics ($R^2$, MAE, and Spearman's $\rho$) for several source-target site combinations, highlighting its effectiveness in tackling multi-source DA with predictive shifts in EEG data analysis. Our method has the potential to combine the advantages of mixed-effects modeling with machine learning for biomedical applications of EEG, such as multicenter clinical trials.

♻ ☆ FLAME: Adaptive and Reactive Concept Drift Mitigation for Federated Learning Deployments

Ioannis Mavromatis, Stefano De Feo, Aftab Khan

This paper presents Federated Learning with Adaptive Monitoring and Elimination (FLAME), a novel solution capable of detecting and mitigating concept drift in Federated Learning (FL) Internet of Things (IoT) environments. Concept drift poses significant challenges for FL models deployed in dynamic and real-world settings. FLAME leverages an FL architecture, considers a real-world FL pipeline, and proves capable of maintaining model performance and accuracy while addressing bandwidth and privacy constraints. Introducing various features and extensions on previous works, FLAME offers a robust solution to concept drift, significantly reducing computational load and communication overhead. Compared to well-known lightweight mitigation methods, FLAME demonstrates superior performance in maintaining high F1 scores and reducing resource utilisation in large-scale IoT deployments, making it a promising approach for real-world applications.

comment: Accepted for Publication at ACM EWSN 2024 - EMERGE Workshop

♻ ☆ Learning Long Range Dependencies on Graphs via Random Walks

Dexiong Chen, Till Hendrik Schulz, Karsten Borgwardt

Message-passing graph neural networks (GNNs) excel at capturing local relationships but struggle with long-range dependencies in graphs. In contrast, graph transformers (GTs) enable global information exchange but often oversimplify the graph structure by representing graphs as sets of fixed-length vectors. This work introduces a novel architecture that overcomes the shortcomings of both approaches by combining the long-range information of random walks with local message passing. By treating random walks as sequences, our architecture leverages recent advances in sequence models to effectively capture long-range dependencies within these walks. Based on this concept, we propose a framework that offers (1) more expressive graph representations through random walk sequences, (2) the ability to utilize any sequence model for capturing long-range dependencies, and (3) the flexibility by integrating various GNN and GT architectures. Our experimental evaluations demonstrate that our approach achieves significant performance improvements on 19 graph and node benchmark datasets, notably outperforming existing methods by up to 13\% on the PascalVoc-SP and COCO-SP datasets. The code is available at https://github.com/BorgwardtLab/NeuralWalker.

♻ ☆ Dynamic Pricing in Securities Lending Market: Application in Revenue Optimization for an Agent Lender Portfolio

Jing Xu, Yung-Cheng Hsu, William Biscarri

Securities lending is an important part of the financial market structure, where agent lenders help long term institutional investors to lend out their securities to short sellers in exchange for a lending fee. Agent lenders within the market seek to optimize revenue by lending out securities at the highest rate possible. Typically, this rate is set by hard-coded business rules or standard supervised machine learning models. These approaches are often difficult to scale and are not adaptive to changing market conditions. Unlike a traditional stock exchange with a centralized limit order book, the securities lending market is organized similarly to an e-commerce marketplace, where agent lenders and borrowers can transact at any agreed price in a bilateral fashion. This similarity suggests that the use of typical methods for addressing dynamic pricing problems in e-commerce could be effective in the securities lending market. We show that existing contextual bandit frameworks can be successfully utilized in the securities lending market. Using offline evaluation on real historical data, we show that the contextual bandit approach can consistently outperform typical approaches by at least 15% in terms of total revenue generated.

comment: 7 pages, 8 figures

♻ ☆ I Bet You Did Not Mean That: Testing Semantic Importance via Betting

Jacopo Teneggi, Jeremias Sulam

Recent works have extended notions of feature importance to semantic concepts that are inherently interpretable to the users interacting with a black-box predictive model. Yet, precise statistical guarantees, such as false positive rate and false discovery rate control, are needed to communicate findings transparently and to avoid unintended consequences in real-world scenarios. In this paper, we formalize the global (i.e., over a population) and local (i.e., for a sample) statistical importance of semantic concepts for the predictions of opaque models by means of conditional independence, which allows for rigorous testing. We use recent ideas of sequential kernelized independence testing (SKIT) to induce a rank of importance across concepts, and showcase the effectiveness and flexibility of our framework on synthetic datasets as well as on image classification tasks using several and diverse vision-language models.

♻ ☆ When Can Transformers Count to n?

Gilad Yehudai, Haim Kaplan, Asma Ghandeharioun, Mor Geva, Amir Globerson

♻ ☆ softmax is not enough (for sharp out-of-distribution)

Petar Veličković, Christos Perivolaropoulos, Federico Barbero, Razvan Pascanu

A key property of reasoning systems is the ability to make sharp decisions on their input data. For contemporary AI systems, a key carrier of sharp behaviour is the softmax function, with its capability to perform differentiable query-key lookups. It is a common belief that the predictive power of networks leveraging softmax arises from "circuits" which sharply perform certain kinds of computations consistently across many diverse inputs. However, for these circuits to be robust, they would need to generalise well to arbitrary valid inputs. In this paper, we dispel this myth: even for tasks as simple as finding the maximum key, any learned circuitry must disperse as the number of items grows at test time. We attribute this to a fundamental limitation of the softmax function to robustly approximate sharp functions, prove this phenomenon theoretically, and propose adaptive temperature as an ad-hoc technique for improving the sharpness of softmax at inference time.

comment: Comments welcome. 15 pages, 7 figures

♻ ☆ On Bits and Bandits: Quantifying the Regret-Information Trade-off

Itai Shufaro, Nadav Merlis, Nir Weinberger, Shie Mannor

In many sequential decision problems, an agent performs a repeated task. He then suffers regret and obtains information that he may use in the following rounds. However, sometimes the agent may also obtain information and avoid suffering regret by querying external sources. We study the trade-off between the information an agent accumulates and the regret it suffers. We invoke information-theoretic methods for obtaining regret lower bounds, that also allow us to easily re-derive several known lower bounds. We introduce the first Bayesian regret lower bounds that depend on the information an agent accumulates. We also prove regret upper bounds using the amount of information the agent accumulates. These bounds show that information measured in bits, can be traded off for regret, measured in reward. Finally, we demonstrate the utility of these bounds in improving the performance of a question-answering task with large language models, allowing us to obtain valuable insights.

♻ ☆ Neural Collaborative Filtering to Detect Anomalies in Human Semantic Trajectories SP

Yueyang Liu, Lance Kennedy, Hossein Amiri, Andreas Züfle

comment: Accepted for publication in the 1st ACM SIGSPATIAL International Workshop on Geospatial Anomaly Detection (GeoAnomalies'24)

♻ ☆ Spectrum Extraction and Clipping for Implicitly Linear Layers

Ali Ebrahimpour Boroojeny, Matus Telgarsky, Hari Sundaram

♻ ☆ Flat Posterior Does Matter For Bayesian Model Averaging

Sungjun Lim, Jeyoon Yeom, Sooyon Kim, Hoyoon Byun, Jinho Kang, Yohan Jung, Jiyoung Jung, Kyungwoo Song

Bayesian neural network (BNN) approximates the posterior distribution of model parameters and utilizes the posterior for prediction via Bayesian Model Averaging (BMA). The quality of the posterior approximation is critical for achieving accurate and robust predictions. It is known that flatness in the loss landscape is strongly associated with generalization performance, and it necessitates consideration to improve the quality of the posterior approximation. In this work, we empirically demonstrate that BNNs often struggle to capture the flatness. Moreover, we provide both experimental and theoretical evidence showing that BMA can be ineffective without ensuring flatness. To address this, we propose Sharpness-Aware Bayesian Model Averaging (SA-BMA), a novel optimizer that seeks flat posteriors by calculating divergence in the parameter space. SA-BMA aligns with the intrinsic nature of BNN and the generalized version of existing sharpness-aware optimizers for DNN. In addition, we suggest a Bayesian Transfer Learning scheme to efficiently leverage pre-trained DNN. We validate the efficacy of SA-BMA in enhancing generalization performance in few-shot classification and distribution shift by ensuring flat posterior.

♻ ☆ TabDDPM: Modelling Tabular Data with Diffusion Models

Akim Kotelnikov, Dmitry Baranchuk, Ivan Rubachev, Artem Babenko

Denoising diffusion probabilistic models are currently becoming the leading paradigm of generative modeling for many important data modalities. Being the most prevalent in the computer vision community, diffusion models have also recently gained some attention in other domains, including speech, NLP, and graph-like data. In this work, we investigate if the framework of diffusion models can be advantageous for general tabular problems, where datapoints are typically represented by vectors of heterogeneous features. The inherent heterogeneity of tabular data makes it quite challenging for accurate modeling, since the individual features can be of completely different nature, i.e., some of them can be continuous and some of them can be discrete. To address such data types, we introduce TabDDPM -- a diffusion model that can be universally applied to any tabular dataset and handles any type of feature. We extensively evaluate TabDDPM on a wide set of benchmarks and demonstrate its superiority over existing GAN/VAE alternatives, which is consistent with the advantage of diffusion models in other fields. Additionally, we show that TabDDPM is eligible for privacy-oriented setups, where the original datapoints cannot be publicly shared.

comment: code https://github.com/yandex-research/tab-ddpm

♻ ☆ Reconstruct Your Previous Conversations! Comprehensively Investigating Privacy Leakage Risks in Conversations with GPT Models EMNLP 2024

Junjie Chu, Zeyang Sha, Michael Backes, Yang Zhang

comment: Accepted in EMNLP 2024. 14 pages, 10 figures

♻ ☆ Visual Question Decomposition on Multimodal Large Language Models EMNLP2024

Haowei Zhang, Jianzhe Liu, Zhen Han, Shuo Chen, Bailan He, Volker Tresp, Zhiqiang Xu, Jindong Gu

comment: Accepted to EMNLP2024 Findings

♻ ☆ Uncertainty-Aware Decision Transformer for Stochastic Driving Environments

Zenan Li, Fan Nie, Qiao Sun, Fang Da, Hang Zhao

Offline Reinforcement Learning (RL) enables policy learning without active interactions, making it especially appealing for self-driving tasks. Recent successes of Transformers inspire casting offline RL as sequence modeling, which, however, fails in stochastic environments with incorrect assumptions that identical actions can consistently achieve the same goal. In this paper, we introduce an UNcertainty-awaRE deciSion Transformer (UNREST) for planning in stochastic driving environments without introducing additional transition or complex generative models. Specifically, UNREST estimates uncertainties by conditional mutual information between transitions and returns. Discovering 'uncertainty accumulation' and 'temporal locality' properties of driving environments, we replace the global returns in decision transformers with truncated returns less affected by environments to learn from actual outcomes of actions rather than environment transitions. We also dynamically evaluate uncertainty at inference for cautious planning. Extensive experiments demonstrate UNREST's superior performance in various driving scenarios and the power of our uncertainty estimation strategy.

♻ ☆ Residual Stream Analysis with Multi-Layer SAEs

Tim Lawson, Lucy Farnik, Conor Houghton, Laurence Aitchison

comment: 34 pages, 26 figures

♻ ☆ Machine listening in a neonatal intensive care unit

Modan Tailleur, Vincent Lostanlen, Jean-Philippe Rivière, Pierre Aumond

Oxygenators, alarm devices, and footsteps are some of the most common sound sources in a hospital. Detecting them has scientific value for environmental psychology but comes with challenges of its own: namely, privacy preservation and limited labeled data. In this paper, we address these two challenges via a combination of edge computing and cloud computing. For privacy preservation, we have designed an acoustic sensor which computes third-octave spectrograms on the fly instead of recording audio waveforms. For sample-efficient machine learning, we have repurposed a pretrained audio neural network (PANN) via spectral transcoding and label space adaptation. A small-scale study in a neonatological intensive care unit (NICU) confirms that the time series of detected events align with another modality of measurement: i.e., electronic badges for parents and healthcare professionals. Hence, this paper demonstrates the feasibility of polyphonic machine listening in a hospital ward while guaranteeing privacy by design.

♻ ☆ Efficient, Multimodal, and Derivative-Free Bayesian Inference With Fisher-Rao Gradient Flows

Yifan Chen, Daniel Zhengyu Huang, Jiaoyang Huang, Sebastian Reich, Andrew M. Stuart

In this paper, we study efficient approximate sampling for probability distributions known up to normalization constants. We specifically focus on a problem class arising in Bayesian inference for large-scale inverse problems in science and engineering applications. The computational challenges we address with the proposed methodology are: (i) the need for repeated evaluations of expensive forward models; (ii) the potential existence of multiple modes; and (iii) the fact that gradient of, or adjoint solver for, the forward model might not be feasible. While existing Bayesian inference methods meet some of these challenges individually, we propose a framework that tackles all three systematically. Our approach builds upon the Fisher-Rao gradient flow in probability space, yielding a dynamical system for probability densities that converges towards the target distribution at a uniform exponential rate. This rapid convergence is advantageous for the computational burden outlined in (i). We apply Gaussian mixture approximations with operator splitting techniques to simulate the flow numerically; the resulting approximation can capture multiple modes thus addressing (ii). Furthermore, we employ the Kalman methodology to facilitate a derivative-free update of these Gaussian components and their respective weights, addressing the issue in (iii). The proposed methodology results in an efficient derivative-free sampler flexible enough to handle multi-modal distributions: Gaussian Mixture Kalman Inversion (GMKI). The effectiveness of GMKI is demonstrated both theoretically and numerically in several experiments with multimodal target distributions, including proof-of-concept and two-dimensional examples, as well as a large-scale application: recovering the Navier-Stokes initial condition from solution data at positive times.

comment: 42 pages, 10 figures

♻ ☆ An active learning method for solving competitive multi-agent decision-making and control problems

Filippo Fabiani, Alberto Bemporad

To identify a stationary action profile for a population of competitive agents, each executing private strategies, we introduce a novel active-learning scheme where a centralized external observer (or entity) can probe the agents' reactions and recursively update simple local parametric estimates of the action-reaction mappings. Under very general working assumptions (not even assuming that a stationary profile exists), sufficient conditions are established to assess the asymptotic properties of the proposed active learning methodology so that, if the parameters characterizing the action-reaction mappings converge, a stationary action profile is achieved. Such conditions hence act also as certificates for the existence of such a profile. Extensive numerical simulations involving typical competitive multi-agent control and decision-making problems illustrate the practical effectiveness of the proposed learning-based approach.

comment: Python package available at https://github.com/bemporad/gnep-learn

♻ ☆ Generalized Consistency Trajectory Models for Image Manipulation

Beomsu Kim, Jaemin Kim, Jeongsol Kim, Jong Chul Ye

♻ ☆ Multi-agent reinforcement learning using echo-state network and its application to pedestrian dynamics

Hisato Komatsu

In recent years, simulations of pedestrians using the multi-agent reinforcement learning (MARL) have been studied. This study considered the roads on a grid-world environment, and implemented pedestrians as MARL agents using an echo-state network and the least squares policy iteration method. Under this environment, the ability of these agents to learn to move forward by avoiding other agents was investigated. Specifically, we considered two types of tasks: the choice between a narrow direct route and a broad detour, and the bidirectional pedestrian flow in a corridor. The simulations results indicated that the learning was successful when the density of the agents was not that high.

comment: 25 pages, 19 figures

♻ ☆ Learning Contrastive Feature Representations for Facial Action Unit Detection

Ziqiao Shang, Bin Liu, Fengmao Lv, Fei Teng, Tianrui Li

comment: 13 pages, 17 figures, submitted to IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

♻ ☆ QMP: Q-switch Mixture of Policies for Multi-Task Behavior Sharing

Grace Zhang, Ayush Jain, Injune Hwang, Shao-Hua Sun, Joseph J. Lim

Multi-task reinforcement learning (MTRL) aims to learn several tasks simultaneously for better sample efficiency than learning them separately. Traditional methods achieve this by sharing parameters or relabeled data between tasks. In this work, we introduce a new framework for sharing behavioral policies across tasks, which can be used in addition to existing MTRL methods. The key idea is to improve each task's off-policy data collection by employing behaviors from other task policies. Selectively sharing helpful behaviors acquired in one task to collect training data for another task can lead to higher-quality trajectories, leading to more sample-efficient MTRL. Thus, we introduce a simple and principled framework called Q-switch mixture of policies (QMP) that selectively shares behavior between different task policies by using the task's Q-function to evaluate and select useful shareable behaviors. We theoretically analyze how QMP improves the sample efficiency of the underlying RL algorithm. Our experiments show that QMP's behavioral policy sharing provides complementary gains over many popular MTRL algorithms and outperforms alternative ways to share behaviors in various manipulation, locomotion, and navigation environments. Videos are available at https://qmp-mtrl.github.io.

♻ ☆ DALL-M: Context-Aware Clinical Data Augmentation with LLMs

Chihcheng Hsieh, Catarina Moreira, Isabel Blanco Nobre, Sandra Costa Sousa, Chun Ouyang, Margot Brereton, Joaquim Jorge, Jacinto C. Nascimento

X-ray images are vital in medical diagnostics, but their effectiveness is limited without clinical context. Radiologists often find chest X-rays insufficient for diagnosing underlying diseases, necessitating comprehensive clinical features and data integration. We present a novel framework to enhance the clinical context through augmentation techniques with clinical tabular data, thereby improving its applicability and reliability in AI medical diagnostics. We introduce a pioneering approach to clinical data augmentation that employs large language models to generate patient contextual synthetic data. This methodology is crucial for training more robust deep learning models in healthcare. It preserves the integrity of real patient data while enriching the dataset with contextually relevant synthetic features, significantly enhancing model performance. Our methodology, termed DALL-M, uses a three-phase feature generation process: (i)clinical context storage, (ii)expert query generation, and (iii)context-aware feature augmentation. DALL-M generates new, clinically relevant features by synthesizing chest X-ray images and reports. Applied to 799 cases using nine features from the MIMIC-IV dataset, it created an augmented set of 91 features. This is the first work to generate contextual values for patients' X-ray reports. Specifically, we provide (i)the capacity of LLMs to generate contextual synthetic values for existing clinical features and (ii)their ability to create entirely new clinically relevant features. Empirical validation with machine learning models showed significant performance improvements. Incorporating augmented features increased the F1 score by 16.5% and Precision and Recall by approximately 25%. DALL-M addresses a critical gap in clinical data augmentation, offering a robust framework for generating contextually enriched datasets.

comment: we introduce a pioneering approach to clinical data augmentation that employs large language models (LLMs) to generate patient contextual synthetic data. It preserves the integrity of real patient data while enriching the dataset with contextually relevant synthetic features, significantly enhancing model performance

♻ ☆ Classification of All Blood Cell Images using ML and DL Models

Rabia Asghar, Sanjay Kumar, Paul Hynds, Abeera Mahfooz

comment: 15

♻ ☆ AQMLator -- An Auto Quantum Machine Learning E-Platform

Tomasz Rybotycki, Piotr Gawron

A successful Machine Learning (ML) model implementation requires three main components: training dataset, suitable model architecture and training procedure. Given dataset and task, finding an appropriate model might be challenging. AutoML, a branch of ML, focuses on automatic architecture search -- a meta method that aims at moving human from ML system design process. The success of ML and the development of quantum computing (QC) in recent years led to a birth of new fascinating field called Quantum Machine Learning (QML) that, amongst others, incorporates quantum computers into ML models. In this paper we present AQMLator, an Auto Quantum Machine Learning platform that aims to automatically propose and train the quantum layers of an ML model with minimal input from the user. This way, data scientists can bypass the entry barrier for QC and use QML. AQMLator uses standard ML libraries, making it easy to introduce into existing ML pipelines.

comment: 15 pages, 3 figures, links to software in the text

♻ ☆ OptiBench Meets ReSocratic: Measure and Improve LLMs for Optimization Modeling

Zhicheng Yang, Yiwei Wang, Yinya Huang, Zhijiang Guo, Wei Shi, Xiongwei Han, Liang Feng, Linqi Song, Xiaodan Liang, Jing Tang

♻ ☆ AlignedCoT: Prompting Large Language Models via Native-Speaking Demonstrations EMNLP 2024

Zhicheng Yang, Yinya Huang, Jing Xiong, Liang Feng, Xiaodan Liang, Yiwei Wang, Jing Tang

Large Language Models prompting, such as using in-context demonstrations, is a mainstream technique for invoking LLMs to perform high-performance and solid complex reasoning (e.g., mathematical reasoning, commonsense reasoning), and has the potential for further human-machine collaborative scientific findings. However, current LLMs are delicate and elusive in prompt words and styles. And there is an unseen gap between LLM understanding and human-written prompts. This paper introduces Alignedcot, an LLM-acquainted prompting technique that includes proficient ``native-speaking'' in in-context learning for the LLMs. Specifically, it achieves consistent and correct step-wise prompts in zero-shot scenarios by progressively probing, refining, and formatting the LLM chain of thoughts so that free from handcrafted few-shot demonstrations while maintaining the prompt quality. We conduct experiments on mathematical reasoning and commonsense reasoning. We find that LLMs with Alignedcot perform significantly superior to them with human-crafted demonstrations. We further apply Alignedcot for rewriting the GSM8K training set, resulting in a GSM8K-Align dataset. We observe its benefits for retrieval augmented generation. The code and data can be found at https://github.com/yangzhch6/AlignedCoT.

comment: Findings of EMNLP 2024

♻ ☆ Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training NeurIPS 2024

Haoran He, Chenjia Bai, Ling Pan, Weinan Zhang, Bin Zhao, Xuelong Li

comment: Accepted by NeurIPS 2024. 24 pages

♻ ☆ Can Large Language Models Understand Symbolic Graphics Programs?

Zeju Qiu, Weiyang Liu, Haiwen Feng, Zhen Liu, Tim Z. Xiao, Katherine M. Collins, Joshua B. Tenenbaum, Adrian Weller, Michael J. Black, Bernhard Schölkopf

comment: Technical Report v2 (46 pages, 24 figures, project page: https://sgp-bench.github.io/, substantial update from v1)

♻ ☆ Cost Estimation in Unit Commitment Problems Using Simulation-Based Inference

Matthias Pirlet, Adrien Bolland, Gilles Louppe, Damien Ernst

The Unit Commitment (UC) problem is a key optimization task in power systems to forecast the generation schedules of power units over a finite time period by minimizing costs while meeting demand and technical constraints. However, many parameters required by the UC problem are unknown, such as the costs. In this work, we estimate these unknown costs using simulation-based inference on an illustrative UC problem, which provides an approximated posterior distribution of the parameters given observed generation schedules and demands. Our results highlight that the learned posterior distribution effectively captures the underlying distribution of the data, providing a range of possible values for the unknown parameters given a past observation. This posterior allows for the estimation of past costs using observed past generation schedules, enabling operators to better forecast future costs and make more robust generation scheduling forecasts. We present avenues for future research to address overconfidence in posterior estimation, enhance the scalability of the methodology and apply it to more complex UC problems modeling the network constraints and renewable energy sources.

♻ ☆ A Dynamic Model of Performative Human-ML Collaboration: Theory and Empirical Evidence

Tom Sühr, Samira Samadi, Chiara Farronato

Machine learning (ML) models are increasingly used in various applications, from recommendation systems in e-commerce to diagnosis prediction in healthcare. In this paper, we present a novel dynamic framework for thinking about the deployment of ML models in a performative, human-ML collaborative system. In our framework, the introduction of ML recommendations changes the data-generating process of human decisions, which are only a proxy to the ground truth and which are then used to train future versions of the model. We show that this dynamic process in principle can converge to different stable points, i.e. where the ML model and the Human+ML system have the same performance. Some of these stable points are suboptimal with respect to the actual ground truth. As a proof of concept, we conduct an empirical user study with 1,408 participants. In the study, humans solve instances of the knapsack problem with the help of machine learning predictions of varying performance. This is an ideal setting because we can identify the actual ground truth, and evaluate the performance of human decisions supported by ML recommendations. We find that for many levels of ML performance, humans can improve upon the ML predictions. We also find that the improvement could be even higher if humans rationally followed the ML recommendations. Finally, we test whether monetary incentives can increase the quality of human decisions, but we fail to find any positive effect. Using our empirical data to approximate our collaborative system suggests that the learning process would dynamically reach an equilibrium performance that is around 92% of the maximum knapsack value. Our results have practical implications for the deployment of ML models in contexts where human decisions may deviate from the indisputable ground truth.

comment: 10 Pages and appendix

♻ ☆ ProteinBench: A Holistic Evaluation of Protein Foundation Models

Fei Ye, Zaixiang Zheng, Dongyu Xue, Yuning Shen, Lihao Wang, Yiming Ma, Yan Wang, Xinyou Wang, Xiangxin Zhou, Quanquan Gu

Recent years have witnessed a surge in the development of protein foundation models, significantly improving performance in protein prediction and generative tasks ranging from 3D structure prediction and protein design to conformational dynamics. However, the capabilities and limitations associated with these models remain poorly understood due to the absence of a unified evaluation framework. To fill this gap, we introduce ProteinBench, a holistic evaluation framework designed to enhance the transparency of protein foundation models. Our approach consists of three key components: (i) A taxonomic classification of tasks that broadly encompass the main challenges in the protein domain, based on the relationships between different protein modalities; (ii) A multi-metric evaluation approach that assesses performance across four key dimensions: quality, novelty, diversity, and robustness; and (iii) In-depth analyses from various user objectives, providing a holistic view of model performance. Our comprehensive evaluation of protein foundation models reveals several key findings that shed light on their current capabilities and limitations. To promote transparency and facilitate further research, we release the evaluation dataset, code, and a public leaderboard publicly for further analysis and a general modular toolkit. We intend for ProteinBench to be a living benchmark for establishing a standardized, in-depth evaluation framework for protein foundation models, driving their development and application while fostering collaboration within the field.

comment: 30 pages, 2 figures and 15 tables

♻ ☆ Efficient Shield Synthesis via State-Space Transformation

Asger Horn Brorholt, Andreas Holck Høeg-Petersen, Kim Guldstrand Larsen, Christian Schilling

We consider the problem of synthesizing safety strategies for control systems, also known as shields. Since the state space is infinite, shields are typically computed over a finite-state abstraction, with the most common abstraction being a rectangular grid. However, for many systems, such a grid does not align well with the safety property or the system dynamics. That is why a coarse grid is rarely sufficient, but a fine grid is typically computationally infeasible to obtain. In this paper, we show that appropriate state-space transformations can still allow to use a coarse grid at almost no computational overhead. We demonstrate in three case studies that our transformation-based synthesis outperforms a standard synthesis by several orders of magnitude. In the first two case studies, we use domain knowledge to select a suitable transformation. In the third case study, we instead report on results in engineering a transformation without domain knowledge.

♻ ☆ UVIP: Model-Free Approach to Evaluate Reinforcement Learning Algorithms

Ilya Levin, Denis Belomestny, Alexey Naumov, Sergey Samsonov

Policy evaluation is an important instrument for the comparison of different algorithms in Reinforcement Learning (RL). Yet even a precise knowledge of the value function $V^{\pi}$ corresponding to a policy $\pi$ does not provide reliable information on how far is the policy $\pi$ from the optimal one. We present a novel model-free upper value iteration procedure $({\sf UVIP})$ that allows us to estimate the suboptimality gap $V^{\star}(x) - V^{\pi}(x)$ from above and to construct confidence intervals for $V^\star$. Our approach relies on upper bounds to the solution of the Bellman optimality equation via martingale approach. We provide theoretical guarantees for ${\sf UVIP}$ under general assumptions and illustrate its performance on a number of benchmark RL problems.

comment: ICOMP-2024 camera-ready version

Multimedia 5

☆ Art2Mus: Bridging Visual Arts and Music through Cross-Modal Generation ECCV 2024

Ivan Rinaldi, Nicola Fanelli, Giovanna Castellano, Gennaro Vessio

comment: Presented at the AI for Visual Arts (AI4VA) workshop at ECCV 2024

☆ FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

Haokun Chen, Hang Li, Yao Zhang, Gengyuan Zhang, Jinhe Bi, Philip Torr, Jindong Gu, Denis Krompass, Volker Tresp

☆ Attentive-based Multi-level Feature Fusion for Voice Disorder Diagnosis

Lipeng Shen, Yifan Xiong, Dongyue Guo, Wei Mo, Lingyu Yu, Hui Yang, Yi Lin

Voice disorders negatively impact the quality of daily life in various ways. However, accurately recognizing the category of pathological features from raw audio remains a considerable challenge due to the limited dataset. A promising method to handle this issue is extracting multi-level pathological information from speech in a comprehensive manner by fusing features in the latent space. In this paper, a novel framework is designed to explore the way of high-quality feature fusion for effective and generalized detection performance. Specifically, the proposed model follows a two-stage training paradigm: (1) ECAPA-TDNN and Wav2vec 2.0 which have shown remarkable effectiveness in various domains are employed to learn the universal pathological information from raw audio; (2) An attentive fusion module is dedicatedly designed to establish the interaction between pathological features projected by EcapTdnn and Wav2vec 2.0 respectively and guide the multi-layer fusion, the entire model is jointly fine-tuned from pre-trained features by the automatic voice pathology detection task. Finally, comprehensive experiments on the FEMH and SVD datasets demonstrate that the proposed framework outperforms the competitive baselines, and achieves the accuracy of 90.51% and 87.68%.

♻ ☆ Improved Screen Content Coding in VVC Using Soft Context Formation

Hannah Och, Shabhrish Reddy Uddehal, Tilo Strutz, André Kaup

Screen content images typically contain a mix of natural and synthetic image parts. Synthetic sections usually are comprised of uniformly colored areas and repeating colors and patterns. In the VVC standard, these properties are exploited using Intra Block Copy and Palette Mode. In this paper, we show that pixel-wise lossless coding can outperform lossy VVC coding in such areas. We propose an enhanced VVC coding approach for screen content images using the principle of soft context formation. First, the image is separated into two layers in a block-wise manner using a learning-based method with four block features. Synthetic image parts are coded losslessly using soft context formation, the rest with VVC.We modify the available soft context formation coder to incorporate information gained by the decoded VVC layer for improved coding efficiency. Using this approach, we achieve Bjontegaard-Delta-rate gains of 4.98% on the evaluated data sets compared to VVC.

comment: 5 pages, 5 figures, 2 tables

♻ ☆ Lighthouse: A User-Friendly Library for Reproducible Video Moment Retrieval and Highlight Detection EMNLP2024

Taichi Nishimura, Shota Nakada, Hokuto Munakata, Tatsuya Komatsu

comment: accepted at EMNLP2024 - system demonstration track

2024-10-06

Distributed, Parallel, and Cluster Computing 3

☆ CONFINE: Preserving Data Secrecy in Decentralized Process Mining

Valerio Goretti, Davide Basile, Luca Barbaro, Claudio Di Ciccio

In the contemporary business landscape, collaboration across multiple organizations offers a multitude of opportunities, including reduced operational costs, enhanced performance, and accelerated technological advancement. The application of process mining techniques in an inter-organizational setting, exploiting the recorded process event data, enables the coordination of joint effort and allows for a deeper understanding of the business. Nevertheless, considerable concerns pertaining to data confidentiality emerge, as organizations frequently demonstrate a reluctance to expose sensitive data demanded for process mining, due to concerns related to privacy and security risks. The presence of conflicting interests among the parties involved can impede the practice of open data sharing. To address these challenges, we propose our approach and toolset named CONFINE, which we developed with the intent of enabling process mining on process event data from multiple providers while preserving the confidentiality and integrity of the original records. To ensure that the presented interaction protocol steps are secure and that the processed information is hidden from both involved and external actors, our approach is based on a decentralized architecture and consists of trusted applications running in Trusted Execution Environments (TEE). In this demo paper, we provide an overview of the core components and functionalities as well as the specific details of its application.

☆ Multi Armed Bandit Algorithms Based Virtual Machine Allocation Policy for Security in Multi-Tenant Distributed Systems

Pravin Patil, Geetanjali Kale, Tanmay Karmarkar, Ruturaj Ghatage

This work proposes a secure and dynamic VM allocation strategy for multi-tenant distributed systems using the Thompson sampling approach. The method proves more effective and secure compared to epsilon-greedy and upper confidence bound methods, showing lower regret levels.,Initially, VM allocation was static, but the unpredictable nature of attacks necessitated a dynamic approach. Historical VM data was analyzed to understand attack responses, with rewards granted for unsuccessful attacks and reduced for successful ones, influencing regret levels.,The paper introduces a Multi Arm Bandit-based VM allocation policy, utilizing a Weighted Average Ensemble Learning algorithm trained on known attacks and non-attacks. This ensemble approach outperforms traditional algorithms like Logistic Regression, SVM, K Nearest Neighbors, and XGBoost.,For suspicious activity detection, a Stacked Anomaly Detector algorithm is proposed, trained on known non-attacks. This method surpasses existing techniques such as Isolation Forest and PCA-based approaches.,Overall, this paper presents an advanced solution for VM allocation policies, enhancing cloud-based system security through a combination of dynamic allocation, ensemble learning, and anomaly detection techniques.

♻ ☆ Model-guided Fuzzing of Distributed Systems

Ege Berkay Gulcan, Burcu Kulahcioglu Ozkan, Rupak Majumdar, Srinidhi Nagendra

We present a coverage-guided testing algorithm for distributed systems implementations. Our main innovation is the use of an abstract formal model of the system that is used to define coverage. Such abstract models are frequently developed in early phases of protocol design and verification but are infrequently used at testing time. We show that guiding random test generation using model coverage can be effective in covering interesting points in the implementation state space. We have implemented a fuzzer for distributed system implementations and abstract models written in TLA+. Our algorithm shows better coverage over purely random exploration as well as random exploration guided by different notions of scheduler coverage and mutation. In particular, we show consistently higher coverage and detect bugs faster on implementations of distributed consensus protocols such as those in Etcd-raft and RedisRaft. Moreover, we discovered 13 previously unknown bugs in their implementations, four of which could only be detected by model-guided fuzzing.

Multimedia 5

☆ Watermarking Decision Tree Ensembles

Stefano Calzavara, Lorenzo Cazzaro, Donald Gera, Salvatore Orlando

Protecting the intellectual property of machine learning models is a hot topic and many watermarking schemes for deep neural networks have been proposed in the literature. Unfortunately, prior work largely neglected the investigation of watermarking techniques for other types of models, including decision tree ensembles, which are a state-of-the-art model for classification tasks on non-perceptual data. In this paper, we present the first watermarking scheme designed for decision tree ensembles, focusing in particular on random forest models. We discuss watermark creation and verification, presenting a thorough security analysis with respect to possible attacks. We finally perform an experimental evaluation of the proposed scheme, showing excellent results in terms of accuracy and security against the most relevant threats.

comment: 7 pages, 5 figures, 2 tables

☆ UniMuMo: Unified Text, Music and Motion Generation

Han Yang, Kun Su, Yutong Zhang, Jiaben Chen, Kaizhi Qian, Gaowen Liu, Chuang Gan

We introduce UniMuMo, a unified multimodal model capable of taking arbitrary text, music, and motion data as input conditions to generate outputs across all three modalities. To address the lack of time-synchronized data, we align unpaired music and motion data based on rhythmic patterns to leverage existing large-scale music-only and motion-only datasets. By converting music, motion, and text into token-based representation, our model bridges these modalities through a unified encoder-decoder transformer architecture. To support multiple generation tasks within a single framework, we introduce several architectural improvements. We propose encoding motion with a music codebook, mapping motion into the same feature space as music. We introduce a music-motion parallel generation scheme that unifies all music and motion generation tasks into a single transformer decoder architecture with a single training task of music-motion joint generation. Moreover, the model is designed by fine-tuning existing pre-trained single-modality models, significantly reducing computational demands. Extensive experiments demonstrate that UniMuMo achieves competitive results on all unidirectional generation benchmarks across music, motion, and text modalities. Quantitative results are available in the \href{https://hanyangclarence.github.io/unimumo_demo/}{project page}.

☆ Knowledge-Guided Dynamic Modality Attention Fusion Framework for Multimodal Sentiment Analysis EMNLP

Xinyu Feng, Yuming Lin, Lihua He, You Li, Liang Chang, Ya Zhou

Multimodal Sentiment Analysis (MSA) utilizes multimodal data to infer the users' sentiment. Previous methods focus on equally treating the contribution of each modality or statically using text as the dominant modality to conduct interaction, which neglects the situation where each modality may become dominant. In this paper, we propose a Knowledge-Guided Dynamic Modality Attention Fusion Framework (KuDA) for multimodal sentiment analysis. KuDA uses sentiment knowledge to guide the model dynamically selecting the dominant modality and adjusting the contributions of each modality. In addition, with the obtained multimodal representation, the model can further highlight the contribution of dominant modality through the correlation evaluation loss. Extensive experiments on four MSA benchmark datasets indicate that KuDA achieves state-of-the-art performance and is able to adapt to different scenarios of dominant modality.

comment: Accepted to EMNLP Findings 2024

♻ ☆ A Multimedia Framework for Continuum Robots: Systematic, Computational, and Control Perspectives

Po-Yu Hsieh, June-Hao Hou

Continuum robots, which often rely on interdisciplinary and multimedia collaborations, have been increasingly recognized for their potential to revolutionize the field of human-computer interaction (HCI) in varied applications due to their adaptive, responsive, and flexible characteristics. Despite their promises, the lack of an integrated framework poses a significant limitation for both users and developers, resulting in inefficiency and complexity during preliminary developments. Thus, this paper introduces a unified framework for continuum robotic systems that addresses these challenges by integrating system architecture, dynamics computation, and control strategy within a computer-aided design (CAD) platform. The proposed method allows for efficient modeling and quick preview of the robot performance, and thus facilitating iterative design and implementation, with a view to enhancing the quality of robot developments.

comment: 9 pages, 10 figures, 1 table

♻ ☆ To Forget or Not? Towards Practical Knowledge Unlearning for Large Language Models EMNLP 2024

Bozhong Tian, Xiaozhuan Liang, Siyuan Cheng, Qingbin Liu, Mengru Wang, Dianbo Sui, Xi Chen, Huajun Chen, Ningyu Zhang

Large Language Models (LLMs) trained on extensive corpora inevitably retain sensitive data, such as personal privacy information and copyrighted material. Recent advancements in knowledge unlearning involve updating LLM parameters to erase specific knowledge. However, current unlearning paradigms are mired in vague forgetting boundaries, often erasing knowledge indiscriminately. In this work, we introduce KnowUnDo, a benchmark containing copyrighted content and user privacy domains to evaluate if the unlearning process inadvertently erases essential knowledge. Our findings indicate that existing unlearning methods often suffer from excessive unlearning. To address this, we propose a simple yet effective method, MemFlex, which utilizes gradient information to precisely target and unlearn sensitive parameters. Experimental results show that MemFlex is superior to existing methods in both precise knowledge unlearning and general knowledge retaining of LLMs. Code and dataset are released at https://github.com/zjunlp/KnowUnDo.

comment: EMNLP 2024 Findings; Code and dataset are released at https://github.com/zjunlp/KnowUnDo

Hardware Architecture 3

☆ Large Language Model Inference Acceleration: A Comprehensive Hardware Perspective

Jinhao Li, Jiaming Xu, Shan Huang, Yonghua Chen, Wen Li, Jun Liu, Yaoxiu Lian, Jiayi Pan, Li Ding, Hao Zhou, Guohao Dai

Large Language Models (LLMs) have demonstrated remarkable capabilities across various fields, from natural language understanding to text generation. Compared to non-generative LLMs like BERT and DeBERTa, generative LLMs like GPT series and Llama series are currently the main focus due to their superior algorithmic performance. The advancements in generative LLMs are closely intertwined with the development of hardware capabilities. Various hardware platforms exhibit distinct hardware characteristics, which can help improve LLM inference performance. Therefore, this paper comprehensively surveys efficient generative LLM inference on different hardware platforms. First, we provide an overview of the algorithm architecture of mainstream generative LLMs and delve into the inference process. Then, we summarize different optimization methods for different platforms such as CPU, GPU, FPGA, ASIC, and PIM/NDP, and provide inference results for generative LLMs. Furthermore, we perform a qualitative and quantitative comparison of inference performance with batch sizes 1 and 8 on different hardware platforms by considering hardware power consumption, absolute inference speed (tokens/s), and energy efficiency (tokens/J). We compare the performance of the same optimization methods across different hardware platforms, the performance across different hardware platforms, and the performance of different methods on the same hardware platform. This provides a systematic and comprehensive summary of existing inference acceleration work by integrating software optimization methods and hardware platforms, which can point to the future trends and potential developments of generative LLMs and hardware technology for edge-side scenarios.

comment: 43 pages, 15 figures

☆ IMAGine: An In-Memory Accelerated GEMV Engine Overlay

MD Arafat Kabir, Tendayi Kamucheka, Nathaniel Fredricks, Joel Mandebi, Jason Bakos, Miaoqing Huang, David Andrews

Processor-in-Memory (PIM) overlays and new redesigned reconfigurable tile fabrics have been proposed to eliminate the von Neumann bottleneck and enable processing performance to scale with BRAM capacity. The performance of these FPGA-based PIM architectures has been limited due to a reduction of the BRAMs maximum clock frequencies and less than ideal scaling of processing elements with increased BRAM capacity. This paper presents IMAGine, an In-Memory Accelerated GEMV engine, a PIM-array accelerator that clocks at the maximum frequency of the BRAM and scales to 100% of the available BRAMs. Comparative analyses are presented showing execution speeds over existing PIM-based GEMV engines on FPGAs and achieving a 2.65x - 3.2x faster clock. An AMD Alveo U55 implementation is presented that achieves a system clock speed of 737 MHz, providing 64K bit-serial multiply-accumulate (MAC) units for GEMV operation. This establishes IMAGine as the fastest PIM-based GEMV overlay, outperforming even the custom PIM-based FPGA accelerators reported to date. Additionally, it surpasses TPU v1-v2 and Alibaba Hanguang 800 in clock speed while offering an equal or greater number of MAC units.

comment: Accepted in FPL 2024

♻ ☆ Self-Managing DRAM: A Low-Cost Framework for Enabling Autonomous and Efficient in-DRAM Operations MICRO 2024

Hasan Hassan, Ataberk Olgun, A. Giray Yaglikci, Haocong Luo, Onur Mutlu

The memory controller is in charge of managing DRAM maintenance operations (e.g., refresh, RowHammer protection, memory scrubbing) to reliably operate modern DRAM chips. Implementing new maintenance operations often necessitates modifications in the DRAM interface, memory controller, and potentially other system components. Such modifications are only possible with a new DRAM standard, which takes a long time to develop, likely leading to slow progress in the adoption of new architectural techniques in DRAM chips. We propose a new low-cost DRAM architecture, Self-Managing DRAM (SMD), that enables autonomous in-DRAM maintenance operations by transferring the responsibility for controlling maintenance operations from the memory controller to the SMD chip. To enable autonomous maintenance operations, we make a single modification to the DRAM interface, such that an SMD chip rejects memory controller accesses to DRAM regions under maintenance, while allowing memory accesses to others. Thus, SMD enables 1) implementing new in-DRAM maintenance mechanisms (or modifying existing ones) with no further changes in the DRAM interface or other system components, and 2) overlapping the latency of a maintenance operation in one DRAM region with the latency of accessing data in another. We evaluate SMD and show that it 1) can be implemented without adding new pins to the DDRx interface with low latency and area overhead, 2) achieves 4.1% average speedup across 20 four-core memory-intensive workloads over a DDR4-based system/DRAM co-design technique that intelligently parallelizes maintenance operations with memory accesses, and 3) guarantees forward progress for rejected memory accesses. We believe and hope SMD can enable innovations in DRAM architecture to rapidly come to fruition. We open source all SMD source code and data at https://github.com/CMU-SAFARI/SelfManagingDRAM.

comment: Extended version of MICRO 2024 paper titled "Self-Managing DRAM: A Low-Cost Framework for Enabling Autonomous and Efficient DRAM Maintenance Operations''

Programming Languages 1

☆ Efficient Linearizability Monitoring for Sets, Stacks, Queues and Priority Queues

Lee Zheng Han, Umang Mathur

In this paper, we consider the problem of automatically monitoring linearizability. Here, one observes an execution of a concurrent program that interacts with a concurrent object and determines if the execution witnesses the violation of linearizability with respect to the sequential specification of the underlying data structure of the concurrent object. This problem has been extensively studied in the past for read-write registers, and both tight upper and lower bounds have been proposed in this case. While this problem has also been studied for the case of other prominent data structures such as stacks and queues, we find that these results are either not extensive or in some cases incorrect. In this paper, we study the problem under the restriction where values inserted in the data types are distinct (in the execution observed). We then show that under such a restriction, the linearizability problem is solvable in polynomial time for these data types. Beyond theoretical soundness and completeness, the algorithms proposed are empirically proven to outperform all state-of-the-art linearizability monitors.

Computation and Language 8

☆ A Cross-Lingual Meta-Learning Method Based on Domain Adaptation for Speech Emotion Recognition

David-Gabriel Ion, Răzvan-Alexandru Smădu, Dumitru-Clementin Cercel, Florin Pop, Mihaela-Claudia Cercel

Best-performing speech models are trained on large amounts of data in the language they are meant to work for. However, most languages have sparse data, making training models challenging. This shortage of data is even more prevalent in speech emotion recognition. Our work explores the model's performance in limited data, specifically for speech emotion recognition. Meta-learning specializes in improving the few-shot learning. As a result, we employ meta-learning techniques on speech emotion recognition tasks, accent recognition, and person identification. To this end, we propose a series of improvements over the multistage meta-learning method. Unlike other works focusing on smaller models due to the high computational cost of meta-learning algorithms, we take a more practical approach. We incorporate a large pre-trained backbone and a prototypical network, making our methods more feasible and applicable. Our most notable contribution is an improved fine-tuning technique during meta-testing that significantly boosts the performance on out-of-distribution datasets. This result, together with incremental improvements from several other works, helped us achieve accuracy scores of 83.78% and 56.30% for Greek and Romanian speech emotion recognition datasets not included in the training or validation splits in the context of 4-way 5-shot learning.

comment: 16 pages, 1 figure, Accepted by WISE 2024

♻ ☆ Survey of Query-based Text Summarization

Hang Yu, Jiawei Han

Query-based text summarization is an important real world problem that requires to condense the prolix text data into a summary under the guidance of the query information provided by users. The topic has been studied for a long time and there are many existing interesting research related to query-based text summarization. Yet much of the work is not systematically surveyed. This survey aims at summarizing some interesting work in query-based text summarization methods as well as related generic text summarization methods. Not all taxonomies in this paper exist the related work to the best of our knowledge and some analysis will be presented.

♻ ☆ Prompts have evil twins EMNLP 2024

Rimon Melamed, Lucas H. McCabe, Tanay Wakhare, Yejin Kim, H. Howie Huang, Enric Boix-Adsera

We discover that many natural-language prompts can be replaced by corresponding prompts that are unintelligible to humans but that provably elicit similar behavior in language models. We call these prompts "evil twins" because they are obfuscated and uninterpretable (evil), but at the same time mimic the functionality of the original natural-language prompts (twins). Remarkably, evil twins transfer between models. We find these prompts by solving a maximum-likelihood problem which has applications of independent interest.

comment: EMNLP 2024 Main, camera-ready

♻ ☆ An Adversarial Perspective on Machine Unlearning for AI Safety

Jakub Łucki, Boyi Wei, Yangsibo Huang, Peter Henderson, Florian Tramèr, Javier Rando

Large language models are finetuned to refuse questions about hazardous knowledge, but these protections can often be bypassed. Unlearning methods aim at completely removing hazardous capabilities from models and make them inaccessible to adversaries. This work challenges the fundamental differences between unlearning and traditional safety post-training from an adversarial perspective. We demonstrate that existing jailbreak methods, previously reported as ineffective against unlearning, can be successful when applied carefully. Furthermore, we develop a variety of adaptive methods that recover most supposedly unlearned capabilities. For instance, we show that finetuning on 10 unrelated examples or removing specific directions in the activation space can recover most hazardous capabilities for models edited with RMU, a state-of-the-art unlearning method. Our findings challenge the robustness of current unlearning approaches and question their advantages over safety training.

♻ ☆ On Efficient Language and Vision Assistants for Visually-Situated Natural Language Understanding: What Matters in Reading and Reasoning EMNLP 2024

Geewook Kim, Minjoon Seo

Recent advancements in language and vision assistants have showcased impressive capabilities but suffer from a lack of transparency, limiting broader research and reproducibility. While open-source models handle general image tasks effectively, they face challenges with the high computational demands of complex visually-situated text understanding. Such tasks often require increased token inputs and large vision modules to harness high-resolution information. Striking a balance between model size and data importance remains an open question. This study aims to redefine the design of vision-language models by identifying key components and creating efficient models with constrained inference costs. By strategically formulating datasets, optimizing vision modules, and enhancing supervision techniques, we achieve significant improvements in inference throughput while maintaining high performance. Extensive experiments across models ranging from 160M to 13B parameters offer insights into model optimization. We will fully open-source our codebase, models, and datasets at https://github.com/naver-ai/elva.

comment: EMNLP 2024 Main

♻ ☆ Cybench: A Framework for Evaluating Cybersecurity Capabilities and Risks of Language Models

Andy K. Zhang, Neil Perry, Riya Dulepet, Joey Ji, Justin W. Lin, Eliot Jones, Celeste Menders, Gashon Hussein, Samantha Liu, Donovan Jasper, Pura Peetathawatchai, Ari Glenn, Vikram Sivashankar, Daniel Zamoshchin, Leo Glikbarg, Derek Askaryar, Mike Yang, Teddy Zhang, Rishi Alluri, Nathan Tran, Rinnara Sangpisit, Polycarpos Yiorkadjis, Kenny Osele, Gautham Raghupathi, Dan Boneh, Daniel E. Ho, Percy Liang

Language Model (LM) agents for cybersecurity that are capable of autonomously identifying vulnerabilities and executing exploits have the potential to cause real-world impact. Policymakers, model providers, and other researchers in the AI and cybersecurity communities are interested in quantifying the capabilities of such agents to help mitigate cyberrisk and investigate opportunities for penetration testing. Toward that end, we introduce Cybench, a framework for specifying cybersecurity tasks and evaluating agents on those tasks. We include 40 professional-level Capture the Flag (CTF) tasks from 4 distinct CTF competitions, chosen to be recent, meaningful, and spanning a wide range of difficulties. Each task includes its own description, starter files, and is initialized in an environment where an agent can execute bash commands and observe outputs. Since many tasks are beyond the capabilities of existing LM agents, we introduce subtasks for each task, which break down a task into intermediary steps for a more detailed evaluation. To evaluate agent capabilities, we construct a cybersecurity agent and evaluate 8 models: GPT-4o, OpenAI o1-preview, Claude 3 Opus, Claude 3.5 Sonnet, Mixtral 8x22b Instruct, Gemini 1.5 Pro, Llama 3 70B Chat, and Llama 3.1 405B Instruct. Without subtask guidance, agents leveraging Claude 3.5 Sonnet, GPT-4o, OpenAI o1-preview, and Claude 3 Opus successfully solved complete tasks that took human teams up to 11 minutes to solve. In comparison, the most difficult task took human teams 24 hours and 54 minutes to solve. All code and data are publicly available at https://cybench.github.io

comment: 78 pages, 6 figures

♻ ☆ COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities

Zihao He, Minh Duc Chu, Rebecca Dorn, Siyi Guo, Kristina Lerman

Social scientists use surveys to probe the opinions and beliefs of populations, but these methods are slow, costly, and prone to biases. Recent advances in large language models (LLMs) enable the creating of computational representations or "digital twins" of populations that generate human-like responses mimicking the population's language, styles, and attitudes. We introduce Community-Cross-Instruct, an unsupervised framework for aligning LLMs to online communities to elicit their beliefs. Given a corpus of a community's online discussions, Community-Cross-Instruct automatically generates instruction-output pairs by an advanced LLM to (1) finetune a foundational LLM to faithfully represent that community, and (2) evaluate the alignment of the finetuned model to the community. We demonstrate the method's utility in accurately representing political and diet communities on Reddit. Unlike prior methods requiring human-authored instructions, Community-Cross-Instruct generates instructions in a fully unsupervised manner, enhancing scalability and generalization across domains. This work enables cost-effective and automated surveying of diverse online communities.

♻ ☆ Superposed Decoding: Multiple Generations from a Single Autoregressive Inference Pass

Ethan Shen, Alan Fan, Sarah M. Pratt, Jae Sung Park, Matthew Wallingford, Sham M. Kakade, Ari Holtzman, Ranjay Krishna, Ali Farhadi, Aditya Kusupati

Many applications today provide users with multiple auto-complete drafts as they type, including GitHub's code completion, Gmail's smart compose, and Apple's messaging auto-suggestions. Under the hood, language models support this by running an autoregressive inference pass to provide a draft. Consequently, providing $k$ drafts to the user requires running an expensive language model $k$ times. To alleviate the computation cost of running $k$ inference passes, we propose Superposed Decoding, a new decoding algorithm that generates $k$ drafts at the computation cost of one autoregressive inference pass. We achieve this by feeding a superposition of the most recent token embeddings from the $k$ drafts as input to the next decoding step of the language model. At every inference step we combine the $k$ drafts with the top-$k$ tokens to get $k^2$ new drafts and cache the $k$ most likely options, using an n-gram interpolation with minimal compute overhead to filter out incoherent generations. Our experiments show that $k$ drafts from Superposed Decoding are at least as coherent and factual as Nucleus Sampling and Greedy Decoding respectively, while being at least $2.44\times$ faster for $k\ge3$. In a compute-normalized setting, user evaluations demonstrably favor text generated by Superposed Decoding over Nucleus Sampling. Superposed Decoding can also be combined with other decoding strategies, resulting in universal coverage gains when scaling inference time compute. Code and more examples open-sourced at https://github.com/RAIVNLab/SuperposedDecoding.

comment: 23 pages, 16 figures

Computer Vision and Pattern Recognition 11

☆ Multimodal 3D Fusion and In-Situ Learning for Spatially Aware AI

Chengyuan Xu, Radha Kumaran, Noah Stier, Kangyou Yu, Tobias Höllerer

Seamless integration of virtual and physical worlds in augmented reality benefits from the system semantically "understanding" the physical environment. AR research has long focused on the potential of context awareness, demonstrating novel capabilities that leverage the semantics in the 3D environment for various object-level interactions. Meanwhile, the computer vision community has made leaps in neural vision-language understanding to enhance environment perception for autonomous tasks. In this work, we introduce a multimodal 3D object representation that unifies both semantic and linguistic knowledge with the geometric representation, enabling user-guided machine learning involving physical objects. We first present a fast multimodal 3D reconstruction pipeline that brings linguistic understanding to AR by fusing CLIP vision-language features into the environment and object models. We then propose "in-situ" machine learning, which, in conjunction with the multimodal representation, enables new tools and interfaces for users to interact with physical spaces and objects in a spatially and linguistically meaningful manner. We demonstrate the usefulness of the proposed system through two real-world AR applications on Magic Leap 2: a) spatial search in physical environments with natural language and b) an intelligent inventory system that tracks object changes over time. We also make our full implementation and demo data available at (https://github.com/cy-xu/spatially_aware_AI) to encourage further exploration and research in spatially aware AI.

comment: 10 pages, 6 figures, accepted to IEEE ISMAR 2024

☆ AdaptDiff: Cross-Modality Domain Adaptation via Weak Conditional Semantic Diffusion for Retinal Vessel Segmentation

Dewei Hu, Hao Li, Han Liu, Jiacheng Wang, Xing Yao, Daiwei Lu, Ipek Oguz

Deep learning has shown remarkable performance in medical image segmentation. However, despite its promise, deep learning has many challenges in practice due to its inability to effectively transition to unseen domains, caused by the inherent data distribution shift and the lack of manual annotations to guide domain adaptation. To tackle this problem, we present an unsupervised domain adaptation (UDA) method named AdaptDiff that enables a retinal vessel segmentation network trained on fundus photography (FP) to produce satisfactory results on unseen modalities (e.g., OCT-A) without any manual labels. For all our target domains, we first adopt a segmentation model trained on the source domain to create pseudo-labels. With these pseudo-labels, we train a conditional semantic diffusion probabilistic model to represent the target domain distribution. Experimentally, we show that even with low quality pseudo-labels, the diffusion model can still capture the conditional semantic information. Subsequently, we sample on the target domain with binary vessel masks from the source domain to get paired data, i.e., target domain synthetic images conditioned on the binary vessel map. Finally, we fine-tune the pre-trained segmentation network using the synthetic paired data to mitigate the domain gap. We assess the effectiveness of AdaptDiff on seven publicly available datasets across three distinct modalities. Our results demonstrate a significant improvement in segmentation performance across all unseen datasets. Our code is publicly available at https://github.com/DeweiHu/AdaptDiff.

☆ Mode-GS: Monocular Depth Guided Anchored 3D Gaussian Splatting for Robust Ground-View Scene Rendering

Yonghan Lee, Jaehoon Choi, Dongki Jung, Jaeseong Yun, Soohyun Ryu, Dinesh Manocha, Suyong Yeon

We present a novel-view rendering algorithm, Mode-GS, for ground-robot trajectory datasets. Our approach is based on using anchored Gaussian splats, which are designed to overcome the limitations of existing 3D Gaussian splatting algorithms. Prior neural rendering methods suffer from severe splat drift due to scene complexity and insufficient multi-view observation, and can fail to fix splats on the true geometry in ground-robot datasets. Our method integrates pixel-aligned anchors from monocular depths and generates Gaussian splats around these anchors using residual-form Gaussian decoders. To address the inherent scale ambiguity of monocular depth, we parameterize anchors with per-view depth-scales and employ scale-consistent depth loss for online scale calibration. Our method results in improved rendering performance, based on PSNR, SSIM, and LPIPS metrics, in ground scenes with free trajectory patterns, and achieves state-of-the-art rendering performance on the R3LIVE odometry dataset and the Tanks and Temples dataset.

☆ Multi-Tiered Self-Contrastive Learning for Medical Microwave Radiometry (MWR) Breast Cancer Detection

Christoforos Galazis, Huiyi Wu, Igor Goryanin

The pursuit of enhanced breast cancer detection and monitoring techniques is a paramount healthcare objective, driving the need for innovative imaging technologies and diagnostic approaches. This study introduces a novel multi-tiered self-contrastive model tailored for the application of microwave radiometry (MWR) breast cancer detection. Our approach encompasses three distinct models: Local-MWR (L-MWR), Regional-MWR (R-MWR), and Global-MWR (G-MWR), each engineered to analyze varying sub-regional comparisons within the breasts. These models are cohesively integrated through the Joint-MWR (J-MWR) network, which leverages the self-contrastive data generated at each analytical level to enhance detection capabilities. Employing a dataset comprising 4,932 cases of female patients, our research showcases the effectiveness of our proposed models. Notably, the J-MWR model distinguishes itself by achieving a Matthews correlation coefficient of 0.74 $\pm$ 0.018, surpassing existing MWR neural networks and contrastive methods. These results highlight the significant potential of self-contrastive learning techniques in improving both the diagnostic accuracy and generalizability of MWR-based breast cancer detection processes. Such advancements hold considerable promise for further investigative and clinical endeavors. The source code is available at: https://github.com/cgalaz01/self_contrastive_mwr

☆ Is What You Ask For What You Get? Investigating Concept Associations in Text-to-Image Models

Salma Abdel Magid, Weiwei Pan, Simon Warchol, Grace Guo, Junsik Kim, Mahia Rahman, Hanspeter Pfister

Text-to-image (T2I) models are increasingly used in impactful real-life applications. As such, there is a growing need to audit these models to ensure that they generate desirable, task-appropriate images. However, systematically inspecting the associations between prompts and generated content in a human-understandable way remains challenging. To address this, we propose \emph{Concept2Concept}, a framework where we characterize conditional distributions of vision language models using interpretable concepts and metrics that can be defined in terms of these concepts. This characterization allows us to use our framework to audit models and prompt-datasets. To demonstrate, we investigate several case studies of conditional distributions of prompts, such as user defined distributions or empirical, real world distributions. Lastly, we implement Concept2Concept as an open-source interactive visualization tool facilitating use by non-technical end-users. Warning: This paper contains discussions of harmful content, including CSAM and NSFW material, which may be disturbing to some readers.

☆ Towards Unsupervised Blind Face Restoration using Diffusion Prior

Tianshu Kuai, Sina Honari, Igor Gilitschenski, Alex Levinshtein

Blind face restoration methods have shown remarkable performance, particularly when trained on large-scale synthetic datasets with supervised learning. These datasets are often generated by simulating low-quality face images with a handcrafted image degradation pipeline. The models trained on such synthetic degradations, however, cannot deal with inputs of unseen degradations. In this paper, we address this issue by using only a set of input images, with unknown degradations and without ground truth targets, to fine-tune a restoration model that learns to map them to clean and contextually consistent outputs. We utilize a pre-trained diffusion model as a generative prior through which we generate high quality images from the natural image distribution while maintaining the input image content through consistency constraints. These generated images are then used as pseudo targets to fine-tune a pre-trained restoration model. Unlike many recent approaches that employ diffusion models at test time, we only do so during training and thus maintain an efficient inference-time performance. Extensive experiments show that the proposed approach can consistently improve the perceptual quality of pre-trained blind face restoration models while maintaining great consistency with the input contents. Our best model also achieves the state-of-the-art results on both synthetic and real-world datasets.

comment: Project page: https://dt-bfr.github.io/

☆ VISTA: A Visual and Textual Attention Dataset for Interpreting Multimodal Models

Harshit, Tolga Tasdizen

The recent developments in deep learning led to the integration of natural language processing (NLP) with computer vision, resulting in powerful integrated Vision and Language Models (VLMs). Despite their remarkable capabilities, these models are frequently regarded as black boxes within the machine learning research community. This raises a critical question: which parts of an image correspond to specific segments of text, and how can we decipher these associations? Understanding these connections is essential for enhancing model transparency, interpretability, and trustworthiness. To answer this question, we present an image-text aligned human visual attention dataset that maps specific associations between image regions and corresponding text segments. We then compare the internal heatmaps generated by VL models with this dataset, allowing us to analyze and better understand the model's decision-making process. This approach aims to enhance model transparency, interpretability, and trustworthiness by providing insights into how these models align visual and linguistic information. We conducted a comprehensive study on text-guided visual saliency detection in these VL models. This study aims to understand how different models prioritize and focus on specific visual elements in response to corresponding text segments, providing deeper insights into their internal mechanisms and improving our ability to interpret their outputs.

♻ ☆ On Efficient Language and Vision Assistants for Visually-Situated Natural Language Understanding: What Matters in Reading and Reasoning EMNLP 2024

Geewook Kim, Minjoon Seo

comment: EMNLP 2024 Main

♻ ☆ Towards a vision foundation model for comprehensive assessment of Cardiac MRI

Athira J Jacob, Indraneel Borgohain, Teodora Chitiboi, Puneet Sharma, Dorin Comaniciu, Daniel Rueckert

Cardiac magnetic resonance imaging (CMR), considered the gold standard for noninvasive cardiac assessment, is a diverse and complex modality requiring a wide variety of image processing tasks for comprehensive assessment of cardiac morphology and function. Advances in deep learning have enabled the development of state-of-the-art (SoTA) models for these tasks. However, model training is challenging due to data and label scarcity, especially in the less common imaging sequences. Moreover, each model is often trained for a specific task, with no connection between related tasks. In this work, we introduce a vision foundation model trained for CMR assessment, that is trained in a self-supervised fashion on 36 million CMR images. We then finetune the model in supervised way for 9 clinical tasks typical to a CMR workflow, across classification, segmentation, landmark localization, and pathology detection. We demonstrate improved accuracy and robustness across all tasks, over a range of available labeled dataset sizes. We also demonstrate improved few-shot learning with fewer labeled samples, a common challenge in medical image analyses. We achieve an out-of-box performance comparable to SoTA for most clinical tasks. The proposed method thus presents a resource-efficient, unified framework for CMR assessment, with the potential to accelerate the development of deep learning-based solutions for image analysis tasks, even with few annotated data available.

comment: 11 pages, 3 figures, 4 tables

♻ ☆ A Cat Is A Cat (Not A Dog!): Unraveling Information Mix-ups in Text-to-Image Encoders through Causal Analysis and Embedding Optimization NeurIPS 2024

Chieh-Yun Chen, Li-Wu Tsao, Chiang Tseng, Hong-Han Shuai

comment: Accepted to NeurIPS 2024

♻ ☆ KISS-Matcher: Fast and Robust Point Cloud Registration Revisited

Hyungtae Lim, Daebeom Kim, Gunhee Shin, Jingnan Shi, Ignacio Vizzo, Hyun Myung, Jaesik Park, Luca Carlone

While global point cloud registration systems have advanced significantly in all aspects, many studies have focused on specific components, such as feature extraction, graph-theoretic pruning, or pose solvers. In this paper, we take a holistic view on the registration problem and develop an open-source and versatile C++ library for point cloud registration, called \textit{KISS-Matcher}. KISS-Matcher combines a novel feature detector, \textit{Faster-PFH}, that improves over the classical fast point feature histogram (FPFH). Moreover, it adopts a $k$-core-based graph-theoretic pruning to reduce the time complexity of rejecting outlier correspondences. Finally, it combines these modules in a complete, user-friendly, and ready-to-use pipeline. As verified by extensive experiments, KISS-Matcher has superior scalability and broad applicability, achieving a substantial speed-up compared to state-of-the-art outlier-robust registration pipelines while preserving accuracy. Our code will be available at \href{https://github.com/MIT-SPARK/KISS-Matcher}{\texttt{https://github.com/MIT-SPARK/KISS-Matcher}}.

comment: 9 pages, 9 figures

2024-10-05

Distributed, Parallel, and Cluster Computing 8

☆ MindFlayer: Efficient Asynchronous Parallel SGD in the Presence of Heterogeneous and Random Worker Compute Times

Artavazd Maranjyan, Omar Shaikh Omar, Peter Richtárik

We study the problem of minimizing the expectation of smooth nonconvex functions with the help of several parallel workers whose role is to compute stochastic gradients. In particular, we focus on the challenging situation where the workers' compute times are arbitrarily heterogeneous and random. In the simpler regime characterized by arbitrarily heterogeneous but deterministic compute times, Tyurin and Richt\'arik (NeurIPS 2023) recently designed the first theoretically optimal asynchronous SGD method, called Rennala SGD, in terms of a novel complexity notion called time complexity. The starting point of our work is the observation that Rennala SGD can have arbitrarily bad performance in the presence of random compute times -- a setting it was not designed to handle. To advance our understanding of stochastic optimization in this challenging regime, we propose a new asynchronous SGD method, for which we coin the name MindFlayer SGD. Our theory and empirical results demonstrate the superiority of MindFlayer SGD over existing baselines, including Rennala SGD, in cases when the noise is heavy tailed.

☆ Slim-ABC: An Optimized Atomic Broadcast Protocol

Nasit S Sony, Xianzhong Ding, Mukesh Singhal

The Byzantine Agreement (BA) problem is a fundamental challenge in distributed systems, focusing on achieving reaching an agreement among parties, some of which may behave maliciously. With the rise of cryptocurrencies, there has been significant interest in developing atomic broadcast protocols, which facilitate agreement on a subset of parties' requests. However, these protocols often come with high communication complexity ($O(ln^2 + \lambda n^3 \log n)$, where $l$ is the bit length of the input, $n$ is the number of parties, and $\lambda$ represents the security parameter bit length). This can lead to inefficiency, especially when the requests across parties exhibit little variation, resulting in unnecessary resource consumption. In this paper, we introduce Slim-ABC, a novel atomic broadcast protocol that eliminates the $O(ln^2 + \lambda n^3 \log n)$ term associated with traditional atomic broadcast protocols. While Slim-ABC reduces the number of accepted requests, it significantly mitigates resource wastage, making it more efficient. The protocol leverages the asynchronous common subset and provable-broadcast mechanisms to achieve a communication complexity of $O(ln^2 + \lambda n^2)$. Despite the trade-off in accepted requests, Slim-ABC maintains robust security by allowing only a fraction ($f+1$) of parties to broadcast requests. We present an extensive efficiency analysis of Slim-ABC, evaluating its performance across key metrics such as message complexity, communication complexity, and time complexity. Additionally, we provide a rigorous security analysis, demonstrating that Slim-ABC satisfies the \textit{agreement}, \textit{validity}, and \textit{totality} properties of the asynchronous common subset protocol.

comment: 21 pages

☆ Advancements in Robotics Process Automation: A Novel Model with Enhanced Empirical Validation and Theoretical Insights

Gokul Pandy, Vivekananda Jayaram, Manjunatha Sughaturu Krishnappa, Balaji Shesharao Ingole, Koushik Kumar Ganeeb, Shenson Joseph

Robotics Process Automation is revolutionizing business operations by significantly enhancing efficiency, productivity, and operational excellence across various industries. This manuscript delivers a comprehensive review of recent advancements in RPA technologies and proposes a novel model designed to elevate RPA capabilities.

comment: 9 pages

☆ Lazy Qubit Reordering for Accelerating Parallel State-Vector-based Quantum Circuit Simulation

Yusuke Teranishi, Shoma Hiraoka, Wataru Mizukami, Masao Okita, Fumihiko Ino

This paper proposes two quantum operation scheduling methods for accelerating parallel state-vector-based quantum circuit simulation using multiple graphics processing units (GPUs). The proposed methods reduce all-to-all communication caused by qubit reordering (QR), which can dominate the overhead of parallel simulation. Our approach eliminates redundant QRs by introducing intentional delays in QR communications such that multiple QRs can be aggregated into a single QR. The delays are carefully introduced based on the principles of time-space tiling, or a cache optimization technique for classical computers, which we use to arrange the execution order of quantum operations. Moreover, we present an extended scheduling method for the hierarchical interconnection of GPU cluster systems to avoid slow inter-node communication. We develop these methods tailored for two primary procedures in variational quantum eigensolver (VQE) simulation: quantum state update (QSU) and expectation value computation (EVC). Experimental validation on 32-GPU executions demonstrates acceleration in QSU and EVC -- up to 54$\times$ and 606$\times$, respectively -- compared to existing methods. Moreover, our extended scheduling method further reduced communication time by up to 15\% in a two-layered interconnected cluster system. Our approach is useful for any quantum circuit simulations, including QSU and/or EVC.

comment: 24 pages, 18 figures

☆ 180 Days After EIP-4844: Will Blob Sharing Solve Dilemma for Small Rollups?

Suhyeon Lee

The introduction of blobs through EIP-4844 has significantly reduced the Data Availability (DA) costs for rollups on Ethereum. However, due to the fixed size of blobs at 128 KB, rollups with low data throughput face a dilemma: they either use blobs inefficiently or decrease the frequency of DA submissions. Blob sharing, where multiple rollups share a single blob, has been proposed as a solution to this problem. This paper examines the effectiveness of blob sharing based on real-world data collected approximately six months after the implementation of EIP-4844. By simulating cost changes using a simple blob sharing format, we demonstrate that blob sharing can substantially improve the costs and DA service quality for small rollups, effectively resolving their dilemma. Notably, we observed cost reductions in USD exceeding 90% for most of the rollups when they cooperate, attributable to the smoothing effect of the blob base fee achieved through blob sharing.

☆ Dispersion on Time-Varying Graphs

Ashish Saxena, Tanvir Kaur, Kaushik Mondal

The dispersion involves the coordination of $k \leq n$ agents on a graph of size $n$ to reach a configuration where at each node at most one agent can be present. It is a well-studied problem. Also, this problem is studied on dynamic graphs with $n$ nodes where at each discrete time step the graph is a connected sub-graph of the complete graph $K_n$. An optimal algorithm is provided assuming global communication and 1-hop visibility of the agents. How this problem pans out on Time-Varying Graphs (TVG) is an open question in the literature. In this work we study this problem on TVG where at each discrete time step the graph is a connected sub-graph of an underlying graph $G$ (known as a footprint) consisting of $n$ nodes. We have the following results even if only one edge from $G$ is missing in the connected sub-graph at any time step and all agents start from a rooted initial configuration. Even with unlimited memory at each agent and 1-hop visibility, it is impossible to solve dispersion for $n$ co-located agents on a TVG in the local communication model. Furthermore, even with unlimited memory at each agent but without 1-hop visibility, it is impossible to achieve dispersion for $n$ co-located agents in the global communication model. From the positive side, the existing algorithm for dispersion on dynamic graphs with the assumptions of global communication and 1-hop visibility works on TVGs as well. This fact and the impossibility results push us to come up with a modified definition of the dispersion problem on TVGs, as one needs to start with more than $n$ agents if the objective is to drop the strong assumptions of global communication and 1-hop visibility. Then, we provide an algorithm to solve the modified dispersion problem on TVG starting with $n+1$ agents with $O(\log n)$ memory per agent while dropping both the assumptions of global communication and 1-hop visibility.

☆ Advancements in Robotics Process Automation: A Novel Model with Enhanced Empirical Validation and Theoretical Insights

Gokul Pandy, Vivekananda Jayaram, Manjunatha Sughaturu Krishnappa, Balaji Shesharao Ingole, Koushik Kumar Ganeeb, Shenson Joseph

comment: 9 pages. European Journal of Computer Science and Information Technology 2024

♻ ☆ A Starting Point for Dynamic Community Detection with Leiden Algorithm

Subhajit Sahu

Real-world graphs often evolve over time, making community or cluster detection a crucial task. In this technical report, we extend three dynamic approaches - Naive-dynamic (ND), Delta-screening (DS), and Dynamic Frontier (DF) - to our multicore implementation of the Leiden algorithm, known for its high-quality community detection. Our experiments, conducted on a server with a 64-core AMD EPYC-7742 processor, show that ND, DS, and DF Leiden achieve average speedups of 1.37x, 1.47x, and 1.98x on large graphs with random batch updates, compared to the Static Leiden algorithm - while scaling at a rate of 1.6x for every doubling of threads. To our knowledge, this is the first attempt to apply dynamic approaches to the Leiden algorithm. We hope these early results pave the way for further development of dynamic approaches for evolving graphs.

comment: 17 pages, 10 figures, 2 tables. arXiv admin note: substantial text overlap with arXiv:2404.19634

Programming Languages 2

♻ ☆ From Code to Correctness: Closing the Last Mile of Code Generation with Hierarchical Debugging

Yuling Shi, Songsong Wang, Chengcheng Wan, Xiaodong Gu

While large language models have made significant strides in code generation, the pass rate of the generated code is bottlenecked on subtle errors, often requiring human intervention to pass tests, especially for complex problems. Existing LLM-based debugging systems treat generated programs as monolithic units, failing to address bugs at multiple levels of granularity, from low-level syntax errors to high-level algorithmic flaws. In this paper, we introduce Multi-Granularity Debugger (MGDebugger), a hierarchical code debugger by isolating, identifying, and resolving bugs at various levels of granularity. MGDebugger decomposes problematic code into a hierarchical tree structure of subfunctions, with each level representing a particular granularity of error. During debugging, it analyzes each subfunction and iteratively resolves bugs in a bottom-up manner. To effectively test each subfunction, we propose an LLM-simulated Python executor, which traces code execution and tracks important variable states to pinpoint errors accurately. Extensive experiments demonstrate that MGDebugger outperforms existing debugging systems, achieving an 18.9% improvement in accuracy over seed generations in HumanEval and a 97.6% repair success rate in HumanEvalFix. Furthermore, MGDebugger effectively fixes bugs across different categories and difficulty levels, demonstrating its robustness and effectiveness.

comment: Code and data available at https://github.com/YerbaPage/MGDebugger

♻ ☆ Comparing large language models and human programmers for generating programming code

Wenpin Hou, Zhicheng Ji

We systematically evaluated the performance of seven large language models in generating programming code using various prompt strategies, programming languages, and task difficulties. GPT-4 substantially outperforms other large language models, including Gemini Ultra and Claude 2. The coding performance of GPT-4 varies considerably with different prompt strategies. In most LeetCode and GeeksforGeeks coding contests evaluated in this study, GPT-4 employing the optimal prompt strategy outperforms 85 percent of human participants. Additionally, GPT-4 demonstrates strong capabilities in translating code between different programming languages and in learning from past errors. The computational efficiency of the code generated by GPT-4 is comparable to that of human programmers. These results suggest that GPT-4 has the potential to serve as a reliable assistant in programming code generation and software development.

Performance 1

♻ ☆ The Power of Training: How Different Neural Network Setups Influence the Energy Demand

Daniel Geißler, Bo Zhou, Mengxi Liu, Sungho Suh, Paul Lukowicz

This work offers a heuristic evaluation of the effects of variations in machine learning training regimes and learning paradigms on the energy consumption of computing, especially HPC hardware with a life-cycle aware perspective. While increasing data availability and innovation in high-performance hardware fuels the training of sophisticated models, it also fosters the fading perception of energy consumption and carbon emission. Therefore, the goal of this work is to raise awareness about the energy impact of general training parameters and processes, from learning rate over batch size to knowledge transfer. Multiple setups with different hyperparameter configurations are evaluated on three different hardware systems. Among many results, we have found out that even with the same model and hardware to reach the same accuracy, improperly set training hyperparameters consume up to 5 times the energy of the optimal setup. We also extensively examined the energy-saving benefits of learning paradigms including recycling knowledge through pretraining and sharing knowledge through multitask training.

Hardware Architecture 1

♻ ☆ Theseus: Exploring Efficient Wafer-Scale Chip Design for Large Language Models

Jingchen Zhu, Chenhao Xue, Yiqi Chen, Zhao Wang, Guangyu Sun

The emergence of the large language model~(LLM) poses an exponential growth of demand for computation throughput, memory capacity, and communication bandwidth. Such a demand growth has significantly surpassed the improvement of corresponding chip designs. With the advancement of fabrication and integration technologies, designers have been developing Wafer-Scale Chips~(WSCs) to scale up and exploit the limits of computation density, memory capacity, and communication bandwidth at the level of a single chip. Existing solutions have demonstrated the significant advantages of WSCs over traditional designs, showing potential to effectively support LLM workloads. Despite the benefits, exploring the early-stage design space of WSCs for LLMs is a crucial yet challenging task due to the enormous and complicated design space, time-consuming evaluation methods, and inefficient exploration strategies. To address these challenges, we propose Theseus, an efficient WSC design space exploration framework for LLMs. We construct the design space of WSCs with various constraints considering the unique characteristics of WSCs. We propose efficient evaluation methodologies for large-scale NoC-based WSCs and introduce multi-fidelity Bayesian optimization to efficiently explore the design space. Evaluation results demonstrate the efficiency of Theseus that the searched Pareto optimal results outperform GPU cluster and existing WSC designs by up to 62.8\%/73.7\% in performance and 38.6\%/42.4\% in power consumption for LLM training, while improving up to 23.2$\times$ and 15.7$\times$ for the performance and power of inference tasks. Furthermore, we conduct case studies to address the design tradeoffs in WSCs and provide insights to facilitate WSC designs for LLMs.

Multimedia 2

☆ AIM 2024 Challenge on Video Super-Resolution Quality Assessment: Methods and Results

Ivan Molodetskikh, Artem Borisov, Dmitriy Vatolin, Radu Timofte, Jianzhao Liu, Tianwu Zhi, Yabin Zhang, Yang Li, Jingwen Xu, Yiting Liao, Qing Luo, Ao-Xiang Zhang, Peng Zhang, Haibo Lei, Linyan Jiang, Yaqing Li, Yuqin Cao, Wei Sun, Weixia Zhang, Yinan Sun, Ziheng Jia, Yuxin Zhu, Xiongkuo Min, Guangtao Zhai, Weihua Luo, Yupeng Z., Hong Y

This paper presents the Video Super-Resolution (SR) Quality Assessment (QA) Challenge that was part of the Advances in Image Manipulation (AIM) workshop, held in conjunction with ECCV 2024. The task of this challenge was to develop an objective QA method for videos upscaled 2x and 4x by modern image- and video-SR algorithms. QA methods were evaluated by comparing their output with aggregate subjective scores collected from >150,000 pairwise votes obtained through crowd-sourced comparisons across 52 SR methods and 1124 upscaled videos. The goal was to advance the state-of-the-art in SR QA, which had proven to be a challenging problem with limited applicability of traditional QA methods. The challenge had 29 registered participants, and 5 teams had submitted their final results, all outperforming the current state-of-the-art. All data, including the private test subset, has been made publicly available on the challenge homepage at https://challenges.videoprocessing.ai/challenges/super-resolution-metrics-challenge.html

comment: 18 pages, 7 figures

♻ ☆ Benchmarking Cross-Domain Audio-Visual Deception Detection

Xiaobao Guo, Zitong Yu, Nithish Muthuchamy Selvaraj, Bingquan Shen, Adams Wai-Kin Kong, Alex C. Kot

Automated deception detection is crucial for assisting humans in accurately assessing truthfulness and identifying deceptive behavior. Conventional contact-based techniques, like polygraph devices, rely on physiological signals to determine the authenticity of an individual's statements. Nevertheless, recent developments in automated deception detection have demonstrated that multimodal features derived from both audio and video modalities may outperform human observers on publicly available datasets. Despite these positive findings, the generalizability of existing audio-visual deception detection approaches across different scenarios remains largely unexplored. To close this gap, we present the first cross-domain audio-visual deception detection benchmark, that enables us to assess how well these methods generalize for use in real-world scenarios. We used widely adopted audio and visual features and different architectures for benchmarking, comparing single-to-single and multi-to-single domain generalization performance. To further exploit the impacts using data from multiple source domains for training, we investigate three types of domain sampling strategies, including domain-simultaneous, domain-alternating, and domain-by-domain for multi-to-single domain generalization evaluation. We also propose an algorithm to enhance the generalization performance by maximizing the gradient inner products between modality encoders, named ``MM-IDGM". Furthermore, we proposed the Attention-Mixer fusion method to improve performance, and we believe that this new cross-domain benchmark will facilitate future research in audio-visual deception detection.

comment: 12 pages

2024-10-04

Hardware Architecture 6

☆ A 9T4R RRAM-Based ACAM for Analogue Template Matching at the Edge

Georgios Papandroulidakis, Shady Agwa, Ahmet Cirakoglu, Themis Prodromakis

The continuous shift of computational bottlenecks to the memory access and data transfer, especially for AI applications, poses the urgent needs of re-engineering the computer architecture fundamentals. Many edge computing applications, like wearable and implantable medical devices, introduce increasingly more challenges to conventional computing systems due to the strict requirements of area and power at the edge. Emerging technologies, like Resistive RAM (RRAM), have shown a promising momentum in developing neuro-inspired analogue computing paradigms capable of achieving high classification capabilities alongside high energy efficiency. In this work, we present a novel RRAM-based Analogue Content Addressable Memory (ACAM) for on-line analogue template matching applications. This ACAM-based template matching architecture aims to achieve energy-efficient classification where low energy is of utmost importance. We are showcasing a highly tuneable novel RRAM-based ACAM pixel implemented using a commercial 180nm CMOS technology and in-house RRAM technology and exhibiting low energy dissipation of approximately 0.036pJ and 0.16pJ for mismatch and match, respectively, at 66MHz with 3V voltage supply. A proof-of-concept system-level implementation based on this novel pixel design is also implemented in 180nm.

☆ Research Directions for Verifiable Crypto-Physically Secure TEEs

Sylvain Bellemare

☆ ORAssistant: A Custom RAG-based Conversational Assistant for OpenROAD

Aviral Kaintura, Palaniappan R, Shui Song Luar, Indira Iyer Almeida

Open-source Electronic Design Automation (EDA) tools are rapidly transforming chip design by addressing key barriers of commercial EDA tools such as complexity, costs, and access. Recent advancements in Large Language Models (LLMs) have further enhanced efficiency in chip design by providing user assistance across a range of tasks like setup, decision-making, and flow automation. This paper introduces ORAssistant, a conversational assistant for OpenROAD, based on Retrieval-Augmented Generation (RAG). ORAssistant aims to improve the user experience for the OpenROAD flow, from RTL-GDSII by providing context-specific responses to common user queries, including installation, command usage, flow setup, and execution, in prose format. Currently, ORAssistant integrates OpenROAD, OpenROAD-flow-scripts, Yosys, OpenSTA, and KLayout. The data model is built from publicly available documentation and GitHub resources. The proposed architecture is scalable, supporting extensions to other open-source tools, operating modes, and LLM models. We use Google Gemini as the base LLM model to build and test ORAssistant. Early evaluation results of the RAG-based model show notable improvements in performance and accuracy compared to non-fine-tuned LLMs.

♻ ☆ Do Not Trust Power Management: A Survey on Internal Energy-based Attacks Circumventing Trusted Execution Environments Security Properties

Gwenn Le Gonidec, Maria Méndez Real, Guillaume Bouffard, Jean-Christophe Prévotet

Over the past few years, several research groups have introduced innovative hardware designs for Trusted Execution Environments (TEEs), aiming to secure applications against potentially compromised privileged software, including the kernel. Since 2015, a new class of software-enabled hardware attacks leveraging energy management mechanisms has emerged. These internal energy-based attacks comprise fault, side-channel and covert channel attacks. Their aim is to bypass TEE security guarantees and expose sensitive information such as cryptographic keys. They have increased in prevalence in the past few years. Popular TEE implementations, such as ARM TrustZone and Intel SGX, incorporate countermeasures against these attacks. However, these countermeasures either hinder the capabilities of the power management mechanisms or have been shown to provide insufficient system protection. This article presents the first comprehensive knowledge survey of these attacks, along with an evaluation of literature countermeasures. We believe that this study will spur further community efforts towards this increasingly important type of attacks.

♻ ☆ Towards Efficient Hyperdimensional Computing Using Photonics

Farbin Fayza, Cansu Demirkiran, Hanning Chen, Che-Kai Liu, Avi Mohan, Hamza Errahmouni, Sanggeon Yun, Mohsen Imani, David Zhang, Darius Bunandar, Ajay Joshi

Over the past few years, silicon photonics-based computing has emerged as a promising alternative to CMOS-based computing for Deep Neural Networks (DNN). Unfortunately, the non-linear operations and the high-precision requirements of DNNs make it extremely challenging to design efficient silicon photonics-based systems for DNN inference and training. Hyperdimensional Computing (HDC) is an emerging, brain-inspired machine learning technique that enjoys several advantages over existing DNNs, including being lightweight, requiring low-precision operands, and being robust to noise introduced by the nonidealities in the hardware. For HDC, computing in-memory (CiM) approaches have been widely used, as CiM reduces the data transfer cost if the operands can fit into the memory. However, inefficient multi-bit operations, high write latency, and low endurance make CiM ill-suited for HDC. On the other hand, the existing electro-photonic DNN accelerators are inefficient for HDC because they are specifically optimized for matrix multiplication in DNNs and consume a lot of power with high-precision data converters. In this paper, we argue that photonic computing and HDC complement each other better than photonic computing and DNNs, or CiM and HDC. We propose PhotoHDC, the first-ever electro-photonic accelerator for HDC training and inference, supporting the basic, record-based, and graph encoding schemes. Evaluating with popular datasets, we show that our accelerator can achieve two to five orders of magnitude lower EDP than the state-of-the-art electro-photonic DNN accelerators for implementing HDC training and inference. PhotoHDC also achieves four orders of magnitude lower energy-delay product than CiM-based accelerators for both HDC training and inference.

♻ ☆ BreakHammer: Enhancing RowHammer Mitigations by Carefully Throttling Suspect Threads MICRO'24

Oğuzhan Canpolat, A. Giray Yağlıkçı, Ataberk Olgun, İsmail Emir Yüksel, Yahya Can Tuğrul, Konstantinos Kanellopoulos, Oğuz Ergin, Onur Mutlu

RowHammer is a major read disturbance mechanism in DRAM where repeatedly accessing (hammering) a row of DRAM cells (DRAM row) induces bitflips in other physically nearby DRAM rows. RowHammer solutions perform preventive actions (e.g., refresh neighbor rows of the hammered row) that mitigate such bitflips to preserve memory isolation, a fundamental building block of security and privacy in modern computing systems. However, preventive actions induce non-negligible memory request latency and system performance overheads as they interfere with memory requests. As shrinking technology node size over DRAM chip generations exacerbates RowHammer, the overheads of RowHammer solutions become prohibitively expensive. As a result, a malicious program can effectively hog the memory system and deny service to benign applications by causing many RowHammer-preventive actions. In this work, we tackle the performance overheads of RowHammer solutions by tracking and throttling the generators of memory accesses that trigger RowHammer solutions. To this end, we propose BreakHammer. BreakHammer 1) observes the time-consuming RowHammer-preventive actions of existing RowHammer mitigation mechanisms, 2) identifies hardware threads that trigger many of these actions, and 3) reduces the memory bandwidth usage of each identified thread. As such, BreakHammer significantly reduces the number of RowHammer-preventive actions performed, thereby improving 1) system performance and DRAM energy, and 2) reducing the maximum slowdown induced on a benign application, with near-zero area overhead. Our extensive evaluations demonstrate that BreakHammer effectively reduces the negative performance, energy, and fairness effects of eight RowHammer mitigation mechanisms. To foster further research we open-source our BreakHammer implementation and scripts at https://github.com/CMU-SAFARI/BreakHammer.

comment: To appear in MICRO'24

Operating Systems 1

♻ ☆ CyberCortex.AI: An AI-based Operating System for Autonomous Robotics and Complex Automation

Sorin Grigorescu, Mihai Zaha

The underlying framework for controlling autonomous robots and complex automation applications are Operating Systems (OS) capable of scheduling perception-and-control tasks, as well as providing real-time data communication to other robotic peers and remote cloud computers. In this paper, we introduce CyberCortex AI, a robotics OS designed to enable heterogeneous AI-based robotics and complex automation applications. CyberCortex AI is a decentralized distributed OS which enables robots to talk to each other, as well as to High Performance Computers (HPC) in the cloud. Sensory and control data from the robots is streamed towards HPC systems with the purpose of training AI algorithms, which are afterwards deployed on the robots. Each functionality of a robot (e.g. sensory data acquisition, path planning, motion control, etc.) is executed within a so-called DataBlock of Filters shared through the internet, where each filter is computed either locally on the robot itself, or remotely on a different robotic system. The data is stored and accessed via a so-called Temporal Addressable Memory (TAM), which acts as a gateway between each filter's input and output. CyberCortex AI has two main components: i) the CyberCortex AI inference system, which is a real-time implementation of the DataBlock running on the robots' embedded hardware, and ii) the CyberCortex AI dojo, which runs on an HPC computer in the cloud, and it is used to design, train and deploy AI algorithms. We present a quantitative and qualitative performance analysis of the proposed approach using two collaborative robotics applications: i) a forest fires prevention system based on an Unitree A1 legged robot and an Anafi Parrot 4K drone, as well as ii) an autonomous driving system which uses CyberCortex AI for collaborative perception and motion control.

Programming Languages 2

☆ Generating Equivalent Representations of Code By A Self-Reflection Approach

Jia Li, Ge Li, Lecheng Wang, Hao Zhu, Zhi Jin

Equivalent Representations (ERs) of code are textual representations that preserve the same semantics as the code itself, e.g., natural language comments and pseudocode. ERs play a critical role in software development and maintenance. However, how to automatically generate ERs of code remains an open challenge. In this paper, we propose a self-reflection approach to generating ERs of code. It enables two Large Language Models (LLMs) to work mutually and produce an ER through a reflection process. Depending on whether constraints on ERs are applied, our approach generates ERs in both open and constrained settings. We conduct a empirical study to generate ERs in two settings and obtain eight findings. (1) Generating ERs in the open setting. In the open setting, we allow LLMs to represent code without any constraints, analyzing the resulting ERs and uncovering five key findings. These findings shed light on how LLMs comprehend syntactic structures, APIs, and numerical computations in code. (2) Generating ERs in the constrained setting. In the constrained setting, we impose constraints on ERs, such as natural language comments, pseudocode, and flowcharts. This allows our approach to address a range of software engineering tasks. Based on our experiments, we have three findings demonstrating that our approach can effectively generate ERs that adhere to specific constraints, thus supporting various software engineering tasks. (3) Future directions. We also discuss potential future research directions, such as deriving intermediate languages for code generation, exploring LLM-friendly requirement descriptions, and further supporting software engineering tasks. We believe that this paper will spark discussions in research communities and inspire many follow-up studies.

♻ ☆ Fair Asynchronous Session Subtyping

Mario Bravetti, Julien Lange, Gianluigi Zavattaro

Session types are widely used as abstractions of asynchronous message passing systems. Refinement for such abstractions is crucial as it allows improvements of a given component without compromising its compatibility with the rest of the system. In the context of session types, the most general notion of refinement is asynchronous session subtyping, which allows message emissions to be anticipated w.r.t. a bounded amount of message consumptions. In this paper we investigate the possibility to anticipate emissions w.r.t. an unbounded amount of consumptions: to this aim we propose to consider fair compliance over asynchronous session types and fair refinement as the relation that preserves it. This allows us to propose a novel variant of session subtyping that leverages the notion of controllability from service contract theory and that is a sound characterisation of fair refinement. In addition, we show that both fair refinement and our novel subtyping are undecidable. We also present a sound algorithm which deals with examples that feature potentially unbounded buffering. Finally, we present an implementation of our algorithm and an empirical evaluation of it on synthetic benchmarks.

Computation and Language 151

☆ Enhance Reasoning by Learning from Mistakes: Peer-Review Knowledge Distillation from Multiple Large Language Models

Zhuochun Li, Yuelyu Ji, Rui Meng, Daqing He

Large language models (LLMs) have exhibited complex reasoning abilities by generating question rationales and demonstrated exceptional performance in natural language processing (NLP) tasks. However, these reasoning capabilities generally emerge in models with tens of billions of parameters, creating significant computational challenges for real-world deployment. Recent research has concentrated on improving open-source smaller models through knowledge distillation (KD) from commercial LLMs. Nevertheless, most of these studies rely solely on the responses from one single LLM as the gold rationale for training. In this paper, we introduce a novel Mistake-Aware Peer-Review Distillation (MAPD) approach: 1) Instead of merely obtaining gold rationales from teachers, our method asks teachers to identify and explain the student's mistakes, providing customized instruction learning data. 2) We design a simulated peer-review process between teacher LLMs, which selects only the generated rationales above the acceptance threshold. This reduces the chance of teachers guessing correctly with flawed rationale, improving instructional data quality. Comprehensive experiments and analysis on mathematical, commonsense, and logical reasoning tasks demonstrate the effectiveness of our method.

comment: 14 pages, 5 figures

☆ Unraveling Cross-Modality Knowledge Conflict in Large Vision-Language Models

Tinghui Zhu, Qin Liu, Fei Wang, Zhengzhong Tu, Muhao Chen

Large Vision-Language Models (LVLMs) have demonstrated impressive capabilities for capturing and reasoning over multimodal inputs. However, these models are prone to parametric knowledge conflicts, which arise from inconsistencies of represented knowledge between their vision and language components. In this paper, we formally define the problem of $\textbf{cross-modality parametric knowledge conflict}$ and present a systematic approach to detect, interpret, and mitigate them. We introduce a pipeline that identifies conflicts between visual and textual answers, showing a persistently high conflict rate across modalities in recent LVLMs regardless of the model size. We further investigate how these conflicts interfere with the inference process and propose a contrastive metric to discern the conflicting samples from the others. Building on these insights, we develop a novel dynamic contrastive decoding method that removes undesirable logits inferred from the less confident modality components based on answer confidence. For models that do not provide logits, we also introduce two prompt-based strategies to mitigate the conflicts. Our methods achieve promising improvements in accuracy on both the ViQuAE and InfoSeek datasets. Specifically, using LLaVA-34B, our proposed dynamic contrastive decoding improves an average accuracy of 2.24%.

comment: Website: https://darthzhu.github.io/cross-modality-knowledge-conflict/

☆ RAFT: Realistic Attacks to Fool Text Detectors EMNLP 2024

James Wang, Ran Li, Junfeng Yang, Chengzhi Mao

Large language models (LLMs) have exhibited remarkable fluency across various tasks. However, their unethical applications, such as disseminating disinformation, have become a growing concern. Although recent works have proposed a number of LLM detection methods, their robustness and reliability remain unclear. In this paper, we present RAFT: a grammar error-free black-box attack against existing LLM detectors. In contrast to previous attacks for language models, our method exploits the transferability of LLM embeddings at the word-level while preserving the original text quality. We leverage an auxiliary embedding to greedily select candidate words to perturb against the target detector. Experiments reveal that our attack effectively compromises all detectors in the study across various domains by up to 99%, and are transferable across source models. Manual human evaluation studies show our attacks are realistic and indistinguishable from original human-written text. We also show that examples generated by RAFT can be used to train adversarially robust detectors. Our work shows that current LLM detectors are not adversarially robust, underscoring the urgent need for more resilient detection mechanisms.

comment: Accepted by EMNLP 2024

☆ Aligning LLMs with Individual Preferences via Interaction

Shujin Wu, May Fung, Cheng Qian, Jeonghwan Kim, Dilek Hakkani-Tur, Heng Ji

As large language models (LLMs) demonstrate increasingly advanced capabilities, aligning their behaviors with human values and preferences becomes crucial for their wide adoption. While previous research focuses on general alignment to principles such as helpfulness, harmlessness, and honesty, the need to account for individual and diverse preferences has been largely overlooked, potentially undermining customized human experiences. To address this gap, we train LLMs that can ''interact to align'', essentially cultivating the meta-skill of LLMs to implicitly infer the unspoken personalized preferences of the current user through multi-turn conversations, and then dynamically align their following behaviors and responses to these inferred preferences. Our approach involves establishing a diverse pool of 3,310 distinct user personas by initially creating seed examples, which are then expanded through iterative self-generation and filtering. Guided by distinct user personas, we leverage multi-LLM collaboration to develop a multi-turn preference dataset containing 3K+ multi-turn conversations in tree structures. Finally, we apply supervised fine-tuning and reinforcement learning to enhance LLMs using this dataset. For evaluation, we establish the ALOE (ALign With CustOmized PrEferences) benchmark, consisting of 100 carefully selected examples and well-designed metrics to measure the customized alignment performance during conversations. Experimental results demonstrate the effectiveness of our method in enabling dynamic, personalized alignment via interaction.

comment: The code and dataset are made public at https://github.com/ShujinWu-0814/ALOE

☆ What Matters for Model Merging at Scale?

Prateek Yadav, Tu Vu, Jonathan Lai, Alexandra Chronopoulou, Manaal Faruqui, Mohit Bansal, Tsendsuren Munkhdalai

Model merging aims to combine multiple expert models into a more capable single model, offering benefits such as reduced storage and serving costs, improved generalization, and support for decentralized model development. Despite its promise, previous studies have primarily focused on merging a few small models. This leaves many unanswered questions about the effect of scaling model size and how it interplays with other key factors -- like the base model quality and number of expert models -- , to affect the merged model's performance. This work systematically evaluates the utility of model merging at scale, examining the impact of these different factors. We experiment with merging fully fine-tuned models using 4 popular merging methods -- Averaging, Task~Arithmetic, Dare, and TIES -- across model sizes ranging from 1B-64B parameters and merging up to 8 different expert models. We evaluate the merged models on both held-in tasks, i.e., the expert's training tasks, and zero-shot generalization to unseen held-out tasks. Our experiments provide several new insights about model merging at scale and the interplay between different factors. First, we find that merging is more effective when experts are created from strong base models, i.e., models with good zero-shot performance. Second, larger models facilitate easier merging. Third merging consistently improves generalization capabilities. Notably, when merging 8 large expert models, the merged models often generalize better compared to the multitask trained models. Fourth, we can better merge more expert models when working with larger models. Fifth, different merging methods behave very similarly at larger scales. Overall, our findings shed light on some interesting properties of model merging while also highlighting some limitations. We hope that this study will serve as a reference point on large-scale merging for upcoming research.

comment: 20 Pages, 7 Figures, 4 Tables

☆ TICKing All the Boxes: Generated Checklists Improve LLM Evaluation and Generation

Jonathan Cook, Tim Rocktäschel, Jakob Foerster, Dennis Aumiller, Alex Wang

Given the widespread adoption and usage of Large Language Models (LLMs), it is crucial to have flexible and interpretable evaluations of their instruction-following ability. Preference judgments between model outputs have become the de facto evaluation standard, despite distilling complex, multi-faceted preferences into a single ranking. Furthermore, as human annotation is slow and costly, LLMs are increasingly used to make these judgments, at the expense of reliability and interpretability. In this work, we propose TICK (Targeted Instruct-evaluation with ChecKlists), a fully automated, interpretable evaluation protocol that structures evaluations with LLM-generated, instruction-specific checklists. We first show that, given an instruction, LLMs can reliably produce high-quality, tailored evaluation checklists that decompose the instruction into a series of YES/NO questions. Each question asks whether a candidate response meets a specific requirement of the instruction. We demonstrate that using TICK leads to a significant increase (46.4% $\to$ 52.2%) in the frequency of exact agreements between LLM judgements and human preferences, as compared to having an LLM directly score an output. We then show that STICK (Self-TICK) can be used to improve generation quality across multiple benchmarks via self-refinement and Best-of-N selection. STICK self-refinement on LiveBench reasoning tasks leads to an absolute gain of $+$7.8%, whilst Best-of-N selection with STICK attains $+$6.3% absolute improvement on the real-world instruction dataset, WildBench. In light of this, structured, multi-faceted self-improvement is shown to be a promising way to further advance LLM capabilities. Finally, by providing LLM-generated checklists to human evaluators tasked with directly scoring LLM responses to WildBench instructions, we notably increase inter-annotator agreement (0.194 $\to$ 0.256).

☆ Efficiently Identifying Watermarked Segments in Mixed-Source Texts

Xuandong Zhao, Chenwen Liao, Yu-Xiang Wang, Lei Li

Text watermarks in large language models (LLMs) are increasingly used to detect synthetic text, mitigating misuse cases like fake news and academic dishonesty. While existing watermarking detection techniques primarily focus on classifying entire documents as watermarked or not, they often neglect the common scenario of identifying individual watermark segments within longer, mixed-source documents. Drawing inspiration from plagiarism detection systems, we propose two novel methods for partial watermark detection. First, we develop a geometry cover detection framework aimed at determining whether there is a watermark segment in long text. Second, we introduce an adaptive online learning algorithm to pinpoint the precise location of watermark segments within the text. Evaluated on three popular watermarking techniques (KGW-Watermark, Unigram-Watermark, and Gumbel-Watermark), our approach achieves high accuracy, significantly outperforming baseline methods. Moreover, our framework is adaptable to other watermarking techniques, offering new insights for precise watermark detection.

☆ Understanding Reasoning in Chain-of-Thought from the Hopfieldian View

Lijie Hu, Liang Liu, Shu Yang, Xin Chen, Zhen Tan, Muhammad Asif Ali, Mengdi Li, Di Wang

Large Language Models have demonstrated remarkable abilities across various tasks, with Chain-of-Thought (CoT) prompting emerging as a key technique to enhance reasoning capabilities. However, existing research primarily focuses on improving performance, lacking a comprehensive framework to explain and understand the fundamental factors behind CoT's success. To bridge this gap, we introduce a novel perspective grounded in the Hopfieldian view of cognition in cognitive neuroscience. We establish a connection between CoT reasoning and key cognitive elements such as stimuli, actions, neural populations, and representation spaces. From our view, we can understand the reasoning process as the movement between these representation spaces. Building on this insight, we develop a method for localizing reasoning errors in the response of CoTs. Moreover, we propose the Representation-of-Thought (RoT) framework, which leverages the robustness of low-dimensional representation spaces to enhance the robustness of the reasoning process in CoTs. Experimental results demonstrate that RoT improves the robustness and interpretability of CoT reasoning while offering fine-grained control over the reasoning process.

comment: 28 pages, a new version of "A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning"

☆ Explicit, Implicit, and Scattered: Revisiting Event Extraction to Capture Complex Arguments EMNLP-2024

Omar Sharif, Joseph Gatto, Madhusudan Basak, Sarah M. Preum

Prior works formulate the extraction of event-specific arguments as a span extraction problem, where event arguments are explicit -- i.e. assumed to be contiguous spans of text in a document. In this study, we revisit this definition of Event Extraction (EE) by introducing two key argument types that cannot be modeled by existing EE frameworks. First, implicit arguments are event arguments which are not explicitly mentioned in the text, but can be inferred through context. Second, scattered arguments are event arguments that are composed of information scattered throughout the text. These two argument types are crucial to elicit the full breadth of information required for proper event modeling. To support the extraction of explicit, implicit, and scattered arguments, we develop a novel dataset, DiscourseEE, which includes 7,464 argument annotations from online health discourse. Notably, 51.2% of the arguments are implicit, and 17.4% are scattered, making DiscourseEE a unique corpus for complex event extraction. Additionally, we formulate argument extraction as a text generation problem to facilitate the extraction of complex argument types. We provide a comprehensive evaluation of state-of-the-art models and highlight critical open challenges in generative event extraction. Our data and codebase are available at https://omar-sharif03.github.io/DiscourseEE.

comment: Accepted in EMNLP-2024 (Main). 21 pages, 8 figures, and 11 tables

☆ Table Question Answering for Low-resourced Indic Languages EMNLP

Vaishali Pal, Evangelos Kanoulas, Andrew Yates, Maarten de Rijke

TableQA is the task of answering questions over tables of structured information, returning individual cells or tables as output. TableQA research has focused primarily on high-resource languages, leaving medium- and low-resource languages with little progress due to scarcity of annotated data and neural models. We address this gap by introducing a fully automatic large-scale tableQA data generation process for low-resource languages with limited budget. We incorporate our data generation method on two Indic languages, Bengali and Hindi, which have no tableQA datasets or models. TableQA models trained on our large-scale datasets outperform state-of-the-art LLMs. We further study the trained models on different aspects, including mathematical reasoning capabilities and zero-shot cross-lingual transfer. Our work is the first on low-resource tableQA focusing on scalable data generation and evaluation procedures. Our proposed data generation method can be applied to any low-resource language with a web presence. We release datasets, models, and code (https://github.com/kolk/Low-Resource-TableQA-Indic-languages).

comment: Accepted at EMNLP,2024

☆ Towards Linguistically-Aware and Language-Independent Tokenization for Large Language Models (LLMs)

Abrar Rahman, Garry Bowlin, Binit Mohanty, Sean McGunigal

This paper presents a comprehensive study on the tokenization techniques employed by state-of-the-art large language models (LLMs) and their implications on the cost and availability of services across different languages, especially low resource languages. The analysis considers multiple LLMs, including GPT-4 (using cl100k_base embeddings), GPT-3 (with p50k_base embeddings), and DaVinci (employing r50k_base embeddings), as well as the widely used BERT base tokenizer. The study evaluates the tokenization variability observed across these models and investigates the challenges of linguistic representation in subword tokenization. The research underscores the importance of fostering linguistically-aware development practices, especially for languages that are traditionally under-resourced. Moreover, this paper introduces case studies that highlight the real-world implications of tokenization choices, particularly in the context of electronic health record (EHR) systems. This research aims to promote generalizable Internationalization (I18N) practices in the development of AI services in this domain and beyond, with a strong emphasis on inclusivity, particularly for languages traditionally underrepresented in AI applications.

☆ Structure-Enhanced Protein Instruction Tuning: Towards General-Purpose Protein Understanding

Wei Wu, Chao Wang, Liyi Chen, Mingze Yin, Yiheng Zhu, Kun Fu, Jieping Ye, Hui Xiong, Zheng Wang

Proteins, as essential biomolecules, play a central role in biological processes, including metabolic reactions and DNA replication. Accurate prediction of their properties and functions is crucial in biological applications. Recent development of protein language models (pLMs) with supervised fine tuning provides a promising solution to this problem. However, the fine-tuned model is tailored for particular downstream prediction task, and achieving general-purpose protein understanding remains a challenge. In this paper, we introduce Structure-Enhanced Protein Instruction Tuning (SEPIT) framework to bridge this gap. Our approach integrates a noval structure-aware module into pLMs to inform them with structural knowledge, and then connects these enhanced pLMs to large language models (LLMs) to generate understanding of proteins. In this framework, we propose a novel two-stage instruction tuning pipeline that first establishes a basic understanding of proteins through caption-based instructions and then refines this understanding using a mixture of experts (MoEs) to learn more complex properties and functional information with the same amount of activated parameters. Moreover, we construct the largest and most comprehensive protein instruction dataset to date, which allows us to train and evaluate the general-purpose protein understanding model. Extensive experimental results on open-ended generation and closed-set answer tasks demonstrate the superior performance of SEPIT over both closed-source general LLMs and open-source LLMs trained with protein knowledge.

☆ Enhancing Data Quality through Simple De-duplication: Navigating Responsible Computational Social Science Research EMNLP 2024

Yida Mu, Mali Jin, Xingyi Song, Nikolaos Aletras

Research in natural language processing (NLP) for Computational Social Science (CSS) heavily relies on data from social media platforms. This data plays a crucial role in the development of models for analysing socio-linguistic phenomena within online communities. In this work, we conduct an in-depth examination of 20 datasets extensively used in NLP for CSS to comprehensively examine data quality. Our analysis reveals that social media datasets exhibit varying levels of data duplication. Consequently, this gives rise to challenges like label inconsistencies and data leakage, compromising the reliability of models. Our findings also suggest that data duplication has an impact on the current claims of state-of-the-art performance, potentially leading to an overestimation of model effectiveness in real-world scenarios. Finally, we propose new protocols and best practices for improving dataset development from social media data and its usage.

comment: Accepted at EMNLP 2024 Main

☆ Re-examining Sexism and Misogyny Classification with Annotator Attitudes EMNLP 2024

Aiqi Jiang, Nikolas Vitsakis, Tanvi Dinkar, Gavin Abercrombie, Ioannis Konstas

Gender-Based Violence (GBV) is an increasing problem online, but existing datasets fail to capture the plurality of possible annotator perspectives or ensure the representation of affected groups. We revisit two important stages in the moderation pipeline for GBV: (1) manual data labelling; and (2) automated classification. For (1), we examine two datasets to investigate the relationship between annotator identities and attitudes and the responses they give to two GBV labelling tasks. To this end, we collect demographic and attitudinal information from crowd-sourced annotators using three validated surveys from Social Psychology. We find that higher Right Wing Authoritarianism scores are associated with a higher propensity to label text as sexist, while for Social Dominance Orientation and Neosexist Attitudes, higher scores are associated with a negative tendency to do so. For (2), we conduct classification experiments using Large Language Models and five prompting strategies, including infusing prompts with annotator information. We find: (i) annotator attitudes affect the ability of classifiers to predict their labels; (ii) including attitudinal information can boost performance when we use well-structured brief annotator descriptions; and (iii) models struggle to reflect the increased complexity and imbalanced classes of the new label sets.

comment: Accepted by EMNLP 2024

☆ MARE: Multi-Aspect Rationale Extractor on Unsupervised Rationale Extraction EMNLP2024

Han Jiang, Junwen Duan, Zhe Qu, Jianxin Wang

Unsupervised rationale extraction aims to extract text snippets to support model predictions without explicit rationale annotation. Researchers have made many efforts to solve this task. Previous works often encode each aspect independently, which may limit their ability to capture meaningful internal correlations between aspects. While there has been significant work on mitigating spurious correlations, our approach focuses on leveraging the beneficial internal correlations to improve multi-aspect rationale extraction. In this paper, we propose a Multi-Aspect Rationale Extractor (MARE) to explain and predict multiple aspects simultaneously. Concretely, we propose a Multi-Aspect Multi-Head Attention (MAMHA) mechanism based on hard deletion to encode multiple text chunks simultaneously. Furthermore, multiple special tokens are prepended in front of the text with each corresponding to one certain aspect. Finally, multi-task training is deployed to reduce the training overhead. Experimental results on two unsupervised rationale extraction benchmarks show that MARE achieves state-of-the-art performance. Ablation studies further demonstrate the effectiveness of our method. Our codes have been available at https://github.com/CSU-NLP-Group/MARE.

comment: Accepted in EMNLP2024(Main) conference

☆ No Need to Talk: Asynchronous Mixture of Language Models

Anastasiia Filippova, Angelos Katharopoulos, David Grangier, Ronan Collobert

We introduce SmallTalk LM, an innovative method for training a mixture of language models in an almost asynchronous manner. Each model of the mixture specializes in distinct parts of the data distribution, without the need of high-bandwidth communication between the nodes training each model. At inference, a lightweight router directs a given sequence to a single expert, according to a short prefix. This inference scheme naturally uses a fraction of the parameters from the overall mixture model. Our experiments on language modeling demonstrate tha SmallTalk LM achieves significantly lower perplexity than dense model baselines for the same total training FLOPs and an almost identical inference cost. Finally, in our downstream evaluations we outperform the dense baseline on $75\%$ of the tasks.

comment: 23 pages

☆ Steering Large Language Models between Code Execution and Textual Reasoning

Yongchao Chen, Harsh Jhamtani, Srinagesh Sharma, Chuchu Fan, Chi Wang

While a lot of recent research focuses on enhancing the textual reasoning capabilities of Large Language Models (LLMs) by optimizing the multi-agent framework or reasoning chains, several benchmark tasks can be solved with 100% success through direct coding, which is more scalable and avoids the computational overhead associated with textual iterating and searching. Textual reasoning has inherent limitations in solving tasks with challenges in math, logics, optimization, and searching, which is unlikely to be solved by simply scaling up the model and data size. The recently released OpenAI GPT Code Interpreter and multi-agent frameworks such as AutoGen have demonstrated remarkable proficiency of integrating code generation and execution to solve complex tasks using LLMs. However, based on our experiments on 7 existing popular methods for steering code/text generation in both single- and multi-turn settings with 14 tasks and 6 types of LLMs (including the new O1-preview), currently there is no optimal method to correctly steer LLMs to write code when needed. We discover some interesting patterns on when models use code vs. textual reasoning with the evolution to task complexity and model sizes, which even result in an astonishingly inverse scaling law. We also discover that results from LLM written code are not always better than using textual reasoning, even if the task could be solved through code. To mitigate the above issues, we propose three methods to better steer LLM code/text generation and achieve a notable improvement. The costs of token lengths and runtime are thoroughly discussed for all the methods. We believe the problem of steering LLM code/text generation is critical for future research and has much space for further improvement. Project Page, Datasets, and Codes are available at https://yongchao98.github.io/CodeSteer/.

comment: 32 pages, 12 figures, 12 tables

☆ CliMedBench: A Large-Scale Chinese Benchmark for Evaluating Medical Large Language Models in Clinical Scenarios

Zetian Ouyang, Yishuai Qiu, Linlin Wang, Gerard de Melo, Ya Zhang, Yanfeng Wang, Liang He

With the proliferation of Large Language Models (LLMs) in diverse domains, there is a particular need for unified evaluation standards in clinical medical scenarios, where models need to be examined very thoroughly. We present CliMedBench, a comprehensive benchmark with 14 expert-guided core clinical scenarios specifically designed to assess the medical ability of LLMs across 7 pivot dimensions. It comprises 33,735 questions derived from real-world medical reports of top-tier tertiary hospitals and authentic examination exercises. The reliability of this benchmark has been confirmed in several ways. Subsequent experiments with existing LLMs have led to the following findings: (i) Chinese medical LLMs underperform on this benchmark, especially where medical reasoning and factual consistency are vital, underscoring the need for advances in clinical knowledge and diagnostic accuracy. (ii) Several general-domain LLMs demonstrate substantial potential in medical clinics, while the limited input capacity of many medical LLMs hinders their practical use. These findings reveal both the strengths and limitations of LLMs in clinical scenarios and offer critical insights for medical research.

comment: accepted by ENMLP-2024

☆ Towards Reproducible LLM Evaluation: Quantifying Uncertainty in LLM Benchmark Scores

Robert E. Blackwell, Jon Barry, Anthony G. Cohn

Large language models (LLMs) are stochastic, and not all models give deterministic answers, even when setting temperature to zero with a fixed random seed. However, few benchmark studies attempt to quantify uncertainty, partly due to the time and cost of repeated experiments. We use benchmarks designed for testing LLMs' capacity to reason about cardinal directions to explore the impact of experimental repeats on mean score and prediction interval. We suggest a simple method for cost-effectively quantifying the uncertainty of a benchmark score and make recommendations concerning reproducible LLM evaluation.

comment: 4 pages, 1 figure

☆ Is Safer Better? The Impact of Guardrails on the Argumentative Strength of LLMs in Hate Speech Countering

Helena Bonaldi, Greta Damo, Nicolás Benjamín Ocampo, Elena Cabrio, Serena Villata, Marco Guerini

The potential effectiveness of counterspeech as a hate speech mitigation strategy is attracting increasing interest in the NLG research community, particularly towards the task of automatically producing it. However, automatically generated responses often lack the argumentative richness which characterises expert-produced counterspeech. In this work, we focus on two aspects of counterspeech generation to produce more cogent responses. First, by investigating the tension between helpfulness and harmlessness of LLMs, we test whether the presence of safety guardrails hinders the quality of the generations. Secondly, we assess whether attacking a specific component of the hate speech results in a more effective argumentative strategy to fight online hate. By conducting an extensive human and automatic evaluation, we show how the presence of safety guardrails can be detrimental also to a task that inherently aims at fostering positive social interactions. Moreover, our results show that attacking a specific component of the hate speech, and in particular its implicit negative stereotype and its hateful parts, leads to higher-quality generations.

comment: To appear in Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (long paper)

☆ Auto-GDA: Automatic Domain Adaptation for Efficient Grounding Verification in Retrieval Augmented Generation

Tobias Leemann, Periklis Petridis, Giuseppe Vietri, Dionysis Manousakas, Aaron Roth, Sergul Aydore

While retrieval augmented generation (RAG) has been shown to enhance factuality of large language model (LLM) outputs, LLMs still suffer from hallucination, generating incorrect or irrelevant information. One common detection strategy involves prompting the LLM again to assess whether its response is grounded in the retrieved evidence, but this approach is costly. Alternatively, lightweight natural language inference (NLI) models for efficient grounding verification can be used at inference time. While existing pre-trained NLI models offer potential solutions, their performance remains subpar compared to larger models on realistic RAG inputs. RAG inputs are more complex than most datasets used for training NLI models and have characteristics specific to the underlying knowledge base, requiring adaptation of the NLI models to a specific target domain. Additionally, the lack of labeled instances in the target domain makes supervised domain adaptation, e.g., through fine-tuning, infeasible. To address these challenges, we introduce Automatic Generative Domain Adaptation (Auto-GDA). Our framework enables unsupervised domain adaptation through synthetic data generation. Unlike previous methods that rely on handcrafted filtering and augmentation strategies, Auto-GDA employs an iterative process to continuously improve the quality of generated samples using weak labels from less efficient teacher models and discrete optimization to select the most promising augmented samples. Experimental results demonstrate the effectiveness of our approach, with models fine-tuned on synthetic data using Auto-GDA often surpassing the performance of the teacher model and reaching the performance level of LLMs at 10 % of their computational cost.

☆ Multi-Dialect Vietnamese: Task, Dataset, Baseline Models and Challenges EMNLP 2024

Nguyen Van Dinh, Thanh Chi Dang, Luan Thanh Nguyen, Kiet Van Nguyen

Vietnamese, a low-resource language, is typically categorized into three primary dialect groups that belong to Northern, Central, and Southern Vietnam. However, each province within these regions exhibits its own distinct pronunciation variations. Despite the existence of various speech recognition datasets, none of them has provided a fine-grained classification of the 63 dialects specific to individual provinces of Vietnam. To address this gap, we introduce Vietnamese Multi-Dialect (ViMD) dataset, a novel comprehensive dataset capturing the rich diversity of 63 provincial dialects spoken across Vietnam. Our dataset comprises 102.56 hours of audio, consisting of approximately 19,000 utterances, and the associated transcripts contain over 1.2 million words. To provide benchmarks and simultaneously demonstrate the challenges of our dataset, we fine-tune state-of-the-art pre-trained models for two downstream tasks: (1) Dialect identification and (2) Speech recognition. The empirical results suggest two implications including the influence of geographical factors on dialects, and the constraints of current approaches in speech recognition tasks involving multi-dialect speech data. Our dataset is available for research purposes.

comment: Main EMNLP 2024

☆ CoCoLoFa: A Dataset of News Comments with Common Logical Fallacies Written by LLM-Assisted Crowds EMNLP 2024

Min-Hsuan Yeh, Ruyuan Wan, Ting-Hao 'Kenneth' Huang

Detecting logical fallacies in texts can help users spot argument flaws, but automating this detection is not easy. Manually annotating fallacies in large-scale, real-world text data to create datasets for developing and validating detection models is costly. This paper introduces CoCoLoFa, the largest known logical fallacy dataset, containing 7,706 comments for 648 news articles, with each comment labeled for fallacy presence and type. We recruited 143 crowd workers to write comments embodying specific fallacy types (e.g., slippery slope) in response to news articles. Recognizing the complexity of this writing task, we built an LLM-powered assistant into the workers' interface to aid in drafting and refining their comments. Experts rated the writing quality and labeling validity of CoCoLoFa as high and reliable. BERT-based models fine-tuned using CoCoLoFa achieved the highest fallacy detection (F1=0.86) and classification (F1=0.87) performance on its test set, outperforming the state-of-the-art LLMs. Our work shows that combining crowdsourcing and LLMs enables us to more effectively construct datasets for complex linguistic phenomena that crowd workers find challenging to produce on their own.

comment: In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024)

☆ How Language Models Prioritize Contextual Grammatical Cues?

Hamidreza Amirzadeh, Afra Alishahi, Hosein Mohebbi

Transformer-based language models have shown an excellent ability to effectively capture and utilize contextual information. Although various analysis techniques have been used to quantify and trace the contribution of single contextual cues to a target task such as subject-verb agreement or coreference resolution, scenarios in which multiple relevant cues are available in the context remain underexplored. In this paper, we investigate how language models handle gender agreement when multiple gender cue words are present, each capable of independently disambiguating a target gender pronoun. We analyze two widely used Transformer-based models: BERT, an encoder-based, and GPT-2, a decoder-based model. Our analysis employs two complementary approaches: context mixing analysis, which tracks information flow within the model, and a variant of activation patching, which measures the impact of cues on the model's prediction. We find that BERT tends to prioritize the first cue in the context to form both the target word representations and the model's prediction, while GPT-2 relies more on the final cue. Our findings reveal striking differences in how encoder-based and decoder-based models prioritize and use contextual information for their predictions.

comment: Accepted to BlackboxNLP 2024

☆ On Uncertainty In Natural Language Processing

Dennis Ulmer

The last decade in deep learning has brought on increasingly capable systems that are deployed on a wide variety of applications. In natural language processing, the field has been transformed by a number of breakthroughs including large language models, which are used in increasingly many user-facing applications. In order to reap the benefits of this technology and reduce potential harms, it is important to quantify the reliability of model predictions and the uncertainties that shroud their development. This thesis studies how uncertainty in natural language processing can be characterized from a linguistic, statistical and neural perspective, and how it can be reduced and quantified through the design of the experimental pipeline. We further explore uncertainty quantification in modeling by theoretically and empirically investigating the effect of inductive model biases in text classification tasks. The corresponding experiments include data for three different languages (Danish, English and Finnish) and tasks as well as a large set of different uncertainty quantification approaches. Additionally, we propose a method for calibrated sampling in natural language generation based on non-exchangeable conformal prediction, which provides tighter token sets with better coverage of the actual continuation. Lastly, we develop an approach to quantify confidence in large black-box language models using auxiliary predictors, where the confidence is predicted from the input to and generated output text of the target model alone.

comment: PhD thesis

☆ Exploring the Benefit of Activation Sparsity in Pre-training ICML 2024

Zhengyan Zhang, Chaojun Xiao, Qiujieli Qin, Yankai Lin, Zhiyuan Zeng, Xu Han, Zhiyuan Liu, Ruobing Xie, Maosong Sun, Jie Zhou

Pre-trained Transformers inherently possess the characteristic of sparse activation, where only a small fraction of the neurons are activated for each token. While sparse activation has been explored through post-training methods, its potential in pre-training remains untapped. In this work, we first study how activation properties change during pre-training. Our examination reveals that Transformers exhibit sparse activation throughout the majority of the pre-training process while the activation correlation keeps evolving as training progresses. Leveraging this observation, we propose Switchable Sparse-Dense Learning (SSD). SSD adaptively switches between the Mixtures-of-Experts (MoE) based sparse training and the conventional dense training during the pre-training process, leveraging the efficiency of sparse training and avoiding the static activation correlation of sparse training. Compared to dense training, SSD achieves comparable performance with identical model size and reduces pre-training costs. Moreover, the models trained with SSD can be directly used as MoE models for sparse inference and achieve the same performance as dense models with up to $2\times$ faster inference speed. Codes are available at https://github.com/thunlp/moefication.

comment: ICML 2024

☆ ToolGen: Unified Tool Retrieval and Calling via Generation

Renxi Wang, Xudong Han, Lei Ji, Shu Wang, Timothy Baldwin, Haonan Li

As large language models (LLMs) advance, their inability to autonomously execute tasks by directly interacting with external tools remains a critical limitation. Traditional methods rely on inputting tool descriptions as context, which is constrained by context length and requires separate, often inefficient, retrieval mechanisms. We introduce ToolGen, a paradigm shift that integrates tool knowledge directly into the LLM's parameters by representing each tool as a unique token. This enables the LLM to generate tool calls and arguments as part of its next token prediction capabilities, seamlessly blending tool invocation with language generation. Our framework allows the LLM to access and utilize a vast amount of tools with no additional retrieval step, significantly enhancing both performance and scalability. Experimental results with over 47,000 tools show that ToolGen not only achieves superior results in both tool retrieval and autonomous task completion but also sets the stage for a new era of AI agents that can adapt to tools across diverse domains. By fundamentally transforming tool retrieval into a generative process, ToolGen paves the way for more versatile, efficient, and autonomous AI systems. ToolGen enables end-to-end tool learning and opens opportunities for integration with other advanced techniques such as chain-of-thought and reinforcement learning, thereby expanding the practical capabilities of LLMs.

☆ A General Framework for Producing Interpretable Semantic Text Embeddings

Yiqun Sun, Qiang Huang, Yixuan Tang, Anthony K. H. Tung, Jun Yu

Semantic text embedding is essential to many tasks in Natural Language Processing (NLP). While black-box models are capable of generating high-quality embeddings, their lack of interpretability limits their use in tasks that demand transparency. Recent approaches have improved interpretability by leveraging domain-expert-crafted or LLM-generated questions, but these methods rely heavily on expert input or well-prompt design, which restricts their generalizability and ability to generate discriminative questions across a wide range of tasks. To address these challenges, we introduce \algo{CQG-MBQA} (Contrastive Question Generation - Multi-task Binary Question Answering), a general framework for producing interpretable semantic text embeddings across diverse tasks. Our framework systematically generates highly discriminative, low cognitive load yes/no questions through the \algo{CQG} method and answers them efficiently with the \algo{MBQA} model, resulting in interpretable embeddings in a cost-effective manner. We validate the effectiveness and interpretability of \algo{CQG-MBQA} through extensive experiments and ablation studies, demonstrating that it delivers embedding quality comparable to many advanced black-box models while maintaining inherently interpretability. Additionally, \algo{CQG-MBQA} outperforms other interpretable text embedding methods across various downstream tasks.

comment: 19 pages, 5 figures, and 9 tables

☆ Images Speak Volumes: User-Centric Assessment of Image Generation for Accessible Communication

Miriam Anschütz, Tringa Sylaj, Georg Groh

Explanatory images play a pivotal role in accessible and easy-to-read (E2R) texts. However, the images available in online databases are not tailored toward the respective texts, and the creation of customized images is expensive. In this large-scale study, we investigated whether text-to-image generation models can close this gap by providing customizable images quickly and easily. We benchmarked seven, four open- and three closed-source, image generation models and provide an extensive evaluation of the resulting images. In addition, we performed a user study with people from the E2R target group to examine whether the images met their requirements. We find that some of the models show remarkable performance, but none of the models are ready to be used at a larger scale without human supervision. Our research is an important step toward facilitating the creation of accessible information for E2R creators and tailoring accessible images to the target group's needs.

comment: To be published at TSAR workshop 2024 (https://tsar-workshop.github.io/)

☆ How Hard is this Test Set? NLI Characterization by Exploiting Training Dynamics EMNLP 2024

Adrian Cosma, Stefan Ruseti, Mihai Dascalu, Cornelia Caragea

Natural Language Inference (NLI) evaluation is crucial for assessing language understanding models; however, popular datasets suffer from systematic spurious correlations that artificially inflate actual model performance. To address this, we propose a method for the automated creation of a challenging test set without relying on the manual construction of artificial and unrealistic examples. We categorize the test set of popular NLI datasets into three difficulty levels by leveraging methods that exploit training dynamics. This categorization significantly reduces spurious correlation measures, with examples labeled as having the highest difficulty showing markedly decreased performance and encompassing more realistic and diverse linguistic phenomena. When our characterization method is applied to the training set, models trained with only a fraction of the data achieve comparable performance to those trained on the full dataset, surpassing other dataset characterization techniques. Our research addresses limitations in NLI dataset construction, providing a more authentic evaluation of model performance with implications for diverse NLU applications.

comment: Accepted at EMNLP 2024 Main Conference

☆ One2set + Large Language Model: Best Partners for Keyphrase Generation EMNLP 2024

Liangying Shao, Liang Zhang, Minlong Peng, Guoqi Ma, Hao Yue, Mingming Sun, Jinsong Su

Keyphrase generation (KPG) aims to automatically generate a collection of phrases representing the core concepts of a given document. The dominant paradigms in KPG include one2seq and one2set. Recently, there has been increasing interest in applying large language models (LLMs) to KPG. Our preliminary experiments reveal that it is challenging for a single model to excel in both recall and precision. Further analysis shows that: 1) the one2set paradigm owns the advantage of high recall, but suffers from improper assignments of supervision signals during training; 2) LLMs are powerful in keyphrase selection, but existing selection methods often make redundant selections. Given these observations, we introduce a generate-then-select framework decomposing KPG into two steps, where we adopt a one2set-based model as generator to produce candidates and then use an LLM as selector to select keyphrases from these candidates. Particularly, we make two important improvements on our generator and selector: 1) we design an Optimal Transport-based assignment strategy to address the above improper assignments; 2) we model the keyphrase selection as a sequence labeling task to alleviate redundant selections. Experimental results on multiple benchmark datasets show that our framework significantly surpasses state-of-the-art models, especially in absent keyphrase prediction.

comment: Accepted by EMNLP 2024 Main Conference

☆ Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation

Xinpeng Wang, Chengzhi Hu, Paul Röttger, Barbara Plank

Training a language model to be both helpful and harmless requires careful calibration of refusal behaviours: Models should refuse to follow malicious instructions or give harmful advice (e.g. "how do I kill someone?"), but they should not refuse safe requests, even if they superficially resemble unsafe ones (e.g. "how do I kill a Python process?"). Avoiding such false refusal, as prior work has shown, is challenging even for highly-capable language models. In this paper, we propose a simple and surgical method for mitigating false refusal in language models via single vector ablation. For a given model, we extract a false refusal vector and show that ablating this vector reduces false refusal rate without negatively impacting model safety and general model capabilities. We also show that our approach can be used for fine-grained calibration of model safety. Our approach is training-free and model-agnostic, making it useful for mitigating the problem of false refusal in current and future language models.

☆ Team MTS @ AutoMin 2021: An Overview of Existing Summarization Approaches and Comparison to Unsupervised Summarization Techniques

Olga Iakovenko, Anna Andreeva, Anna Lapidus, Liana Mikaelyan

Remote communication through video or audio conferences has become more popular than ever because of the worldwide pandemic. These events, therefore, have provoked the development of systems for automatic minuting of spoken language leading to AutoMin 2021 challenge. The following paper illustrates the results of the research that team MTS has carried out while participating in the Automatic Minutes challenge. In particular, in this paper we analyze existing approaches to text and speech summarization, propose an unsupervised summarization technique based on clustering and provide a pipeline that includes an adapted automatic speech recognition block able to run on real-life recordings. The proposed unsupervised technique outperforms pre-trained summarization models on the automatic minuting task with Rouge 1, Rouge 2 and Rouge L values of 0.21, 0.02 and 0.2 on the dev set, with Rouge 1, Rouge 2, Rouge L, Adequacy, Grammatical correctness and Fluency values of 0.180, 0.035, 0.098, 1.857, 2.304, 1.911 on the test set accordingly

comment: First Shared Task on Automatic Minuting at Interspeech 2021

☆ Killing Two Flies with One Stone: An Attempt to Break LLMs Using English->Icelandic Idioms and Proper Names

Bjarki Ármannsson, Hinrik Hafsteinsson, Atli Jasonarson, Steinþór Steingrímsson

This paper presents the submission of the \'Arni Magn\'usson Institute's team to the WMT24 test suite subtask, focusing on idiomatic expressions and proper names for the English->Icelandic translation direction. Intuitively and empirically, idioms and proper names are known to be a significant challenge for modern translation models. We create two different test suites. The first evaluates the competency of MT systems in translating common English idiomatic expressions, as well as testing whether systems can distinguish between those expressions and the same phrases when used in a literal context. The second test suite consists of place names that should be translated into their Icelandic exonyms (and correctly inflected) and pairs of Icelandic names that share a surface form between the male and female variants, so that incorrect translations impact meaning as well as readability. The scores reported are relatively low, especially for idiomatic expressions and place names, and indicate considerable room for improvement.

comment: WMT24 MT Test Suites subtask. 8 pages, 5 tables

☆ Cogs in a Machine, Doing What They're Meant to Do -- The AMI Submission to the WMT24 General Translation Task

Atli Jasonarson, Hinrik Hafsteinsson, Bjarki Ármannsson, Steinþór Steingrímsson

This paper presents the submission of the \'Arni Magnusson Institute's team to the WMT24 General translation task. We work on the English->Icelandic translation direction. Our system comprises four translation models and a grammar correction model. For training our models we carefully curate our datasets, aggressively filtering out sentence pairs that may detrimentally affect the quality of our system's output. Some of our data are collected from human translations and some are synthetically generated. A part of the synthetic data is generated using an LLM, and we find that it increases the translation capability of our system significantly.

comment: WMT24 General Translation Task System Description Paper, 10 pages, 1 figure, 6 tables

☆ Should Cross-Lingual AMR Parsing go Meta? An Empirical Assessment of Meta-Learning and Joint Learning AMR Parsing EMNLP 2024

Jeongwoo Kang, Maximin Coavoux, Cédric Lopez, Didier Schwab

Cross-lingual AMR parsing is the task of predicting AMR graphs in a target language when training data is available only in a source language. Due to the small size of AMR training data and evaluation data, cross-lingual AMR parsing has only been explored in a small set of languages such as English, Spanish, German, Chinese, and Italian. Taking inspiration from Langedijk et al. (2022), who apply meta-learning to tackle cross-lingual syntactic parsing, we investigate the use of meta-learning for cross-lingual AMR parsing. We evaluate our models in $k$-shot scenarios (including 0-shot) and assess their effectiveness in Croatian, Farsi, Korean, Chinese, and French. Notably, Korean and Croatian test sets are developed as part of our work, based on the existing The Little Prince English AMR corpus, and made publicly available. We empirically study our method by comparing it to classical joint learning. Our findings suggest that while the meta-learning model performs slightly better in 0-shot evaluation for certain languages, the performance gain is minimal or absent when $k$ is higher than 0.

comment: to appear in Findings of EMNLP 2024

☆ Generating Equivalent Representations of Code By A Self-Reflection Approach

Jia Li, Ge Li, Lecheng Wang, Hao Zhu, Zhi Jin

☆ Zero-Shot Fact Verification via Natural Logic and Large Language Models EMNLP 2024

Marek Strong, Rami Aly, Andreas Vlachos

The recent development of fact verification systems with natural logic has enhanced their explainability by aligning claims with evidence through set-theoretic operators, providing faithful justifications. Despite these advancements, such systems often rely on a large amount of training data annotated with natural logic. To address this issue, we propose a zero-shot method that utilizes the generalization capabilities of instruction-tuned large language models. To comprehensively assess the zero-shot capabilities of our method and other fact verification systems, we evaluate all models on both artificial and real-world claims, including multilingual datasets. We also compare our method against other fact verification systems in two setups. First, in the zero-shot generalization setup, we demonstrate that our approach outperforms other systems that were not specifically trained on natural logic data, achieving an average accuracy improvement of 8.96 points over the best-performing baseline. Second, in the zero-shot transfer setup, we show that current systems trained on natural logic data do not generalize well to other domains, and our method outperforms these systems across all datasets with real-world claims.

comment: Accepted to EMNLP 2024

☆ Context and System Fusion in Post-ASR Emotion Recognition with Large Language Models

Pavel Stepachev, Pinzhen Chen, Barry Haddow

Large language models (LLMs) have started to play a vital role in modelling speech and text. To explore the best use of context and multiple systems' outputs for post-ASR speech emotion prediction, we study LLM prompting on a recent task named GenSEC. Our techniques include ASR transcript ranking, variable conversation context, and system output fusion. We show that the conversation context has diminishing returns and the metric used to select the transcript for prediction is crucial. Finally, our best submission surpasses the provided baseline by 20% in absolute accuracy.

☆ Comparing zero-shot self-explanations with human rationales in multilingual text classification

Stephanie Brandl, Oliver Eberle

Instruction-tuned LLMs are able to provide an explanation about their output to users by generating self-explanations that do not require gradient computations or the application of possibly complex XAI methods. In this paper, we analyse whether this ability results in a good explanation by evaluating self-explanations in the form of input rationales with respect to their plausibility to humans as well as their faithfulness to models. For this, we apply two text classification tasks: sentiment classification and forced labour detection. Next to English, we further include Danish and Italian translations of the sentiment classification task and compare self-explanations to human annotations for all samples. To allow for direct comparisons, we also compute post-hoc feature attribution, i.e., layer-wise relevance propagation (LRP) and apply this pipeline to 4 LLMs (Llama2, Llama3, Mistral and Mixtral). Our results show that self-explanations align more closely with human annotations compared to LRP, while maintaining a comparable level of faithfulness.

comment: preprint

☆ Five Years of COVID-19 Discourse on Instagram: A Labeled Instagram Dataset of Over Half a Million Posts for Multilingual Sentiment Analysis

Nirmalya Thakur

The work presented in this paper makes three scientific contributions with a specific focus on mining and analysis of COVID-19-related posts on Instagram. First, it presents a multilingual dataset of 500,153 Instagram posts about COVID-19 published between January 2020 and September 2024. This dataset, available at https://dx.doi.org/10.21227/d46p-v480, contains Instagram posts in 161 different languages as well as 535,021 distinct hashtags. After the development of this dataset, multilingual sentiment analysis was performed, which involved classifying each post as positive, negative, or neutral. The results of sentiment analysis are presented as a separate attribute in this dataset. Second, it presents the results of performing sentiment analysis per year from 2020 to 2024. The findings revealed the trends in sentiment related to COVID-19 on Instagram since the beginning of the pandemic. For instance, between 2020 and 2024, the sentiment trends show a notable shift, with positive sentiment decreasing from 38.35% to 28.69%, while neutral sentiment rising from 44.19% to 58.34%. Finally, the paper also presents findings of language-specific sentiment analysis. This analysis highlighted similar and contrasting trends of sentiment across posts published in different languages on Instagram. For instance, out of all English posts, 49.68% were positive, 14.84% were negative, and 35.48% were neutral. In contrast, among Hindi posts, 4.40% were positive, 57.04% were negative, and 38.56% were neutral, reflecting distinct differences in the sentiment distribution between these two languages.

☆ What do Large Language Models Need for Machine Translation Evaluation?

Shenbin Qian, Archchana Sindhujan, Minnie Kabra, Diptesh Kanojia, Constantin Orăsan, Tharindu Ranasinghe, Frédéric Blain

Leveraging large language models (LLMs) for various natural language processing tasks has led to superlative claims about their performance. For the evaluation of machine translation (MT), existing research shows that LLMs are able to achieve results comparable to fine-tuned multilingual pre-trained language models. In this paper, we explore what translation information, such as the source, reference, translation errors and annotation guidelines, is needed for LLMs to evaluate MT quality. In addition, we investigate prompting techniques such as zero-shot, Chain of Thought (CoT) and few-shot prompting for eight language pairs covering high-, medium- and low-resource languages, leveraging varying LLM variants. Our findings indicate the importance of reference translations for an LLM-based evaluation. While larger models do not necessarily fare better, they tend to benefit more from CoT prompting, than smaller models. We also observe that LLMs do not always provide a numerical score when generating evaluations, which poses a question on their reliability for the task. Our work presents a comprehensive analysis for resource-constrained and training-less LLM-based evaluation of machine translation. We release the accrued prompt templates, code and data publicly for reproducibility.

☆ A Multi-task Learning Framework for Evaluating Machine Translation of Emotion-loaded User-generated Content

Shenbin Qian, Constantin Orăsan, Diptesh Kanojia, Félix do Carmo

Machine translation (MT) of user-generated content (UGC) poses unique challenges, including handling slang, emotion, and literary devices like irony and sarcasm. Evaluating the quality of these translations is challenging as current metrics do not focus on these ubiquitous features of UGC. To address this issue, we utilize an existing emotion-related dataset that includes emotion labels and human-annotated translation errors based on Multi-dimensional Quality Metrics. We extend it with sentence-level evaluation scores and word-level labels, leading to a dataset suitable for sentence- and word-level translation evaluation and emotion classification, in a multi-task setting. We propose a new architecture to perform these tasks concurrently, with a novel combined loss function, which integrates different loss heuristics, like the Nash and Aligned losses. Our evaluation compares existing fine-tuning and multi-task learning approaches, assessing generalization with ablative experiments over multiple datasets. Our approach achieves state-of-the-art performance and we present a comprehensive analysis for MT evaluation of UGC.

☆ Adaptive BPE Tokenization for Enhanced Vocabulary Adaptation in Finetuning Pretrained Language Models EMNLP

Gunjan Balde, Soumyadeep Roy, Mainack Mondal, Niloy Ganguly

In this work, we show a fundamental limitation in vocabulary adaptation approaches that use Byte-Pair Encoding (BPE) tokenization scheme for fine-tuning pretrained language models (PLMs) to expert domains. Current approaches trivially append the target domain-specific vocabulary at the end of the PLM vocabulary. This approach leads to a lower priority score and causes sub-optimal tokenization in BPE that iteratively uses merge rules to tokenize a given text. To mitigate this issue, we propose AdaptBPE where the BPE tokenization initialization phase is modified to first perform the longest string matching on the added (target) vocabulary before tokenizing at the character level. We perform an extensive evaluation of AdaptBPE versus the standard BPE over various classification and summarization tasks; AdaptBPE improves by 3.57% (in terms of accuracy) and 1.87% (in terms of Rouge-L), respectively. AdaptBPE for MEDVOC works particularly well when reference summaries have high OOV concentration or are longer in length. We also conduct a human evaluation, revealing that AdaptBPE generates more relevant and more faithful summaries as compared to MEDVOC. We make our codebase publicly available at https://github.com/gb-kgp/adaptbpe.

comment: 11 pages. Accepted at EMNLP Findings 2024 (The 2024 Conference on Empirical Methods in Natural Language Processing)

☆ Towards a Benchmark for Large Language Models for Business Process Management Tasks

Kiran Busch, Henrik Leopold

An increasing number of organizations are deploying Large Language Models (LLMs) for a wide range of tasks. Despite their general utility, LLMs are prone to errors, ranging from inaccuracies to hallucinations. To objectively assess the capabilities of existing LLMs, performance benchmarks are conducted. However, these benchmarks often do not translate to more specific real-world tasks. This paper addresses the gap in benchmarking LLM performance in the Business Process Management (BPM) domain. Currently, no BPM-specific benchmarks exist, creating uncertainty about the suitability of different LLMs for BPM tasks. This paper systematically compares LLM performance on four BPM tasks focusing on small open-source models. The analysis aims to identify task-specific performance variations, compare the effectiveness of open-source versus commercial models, and assess the impact of model size on BPM task performance. This paper provides insights into the practical applications of LLMs in BPM, guiding organizations in selecting appropriate models for their specific needs.

☆ Are Expert-Level Language Models Expert-Level Annotators? NeurIPS 2024

Yu-Min Tseng, Wei-Lin Chen, Chung-Chi Chen, Hsin-Hsi Chen

Data annotation refers to the labeling or tagging of textual data with relevant information. A large body of works have reported positive results on leveraging LLMs as an alternative to human annotators. However, existing studies focus on classic NLP tasks, and the extent to which LLMs as data annotators perform in domains requiring expert knowledge remains underexplored. In this work, we investigate comprehensive approaches across three highly specialized domains and discuss practical suggestions from a cost-effectiveness perspective. To the best of our knowledge, we present the first systematic evaluation of LLMs as expert-level data annotators.

comment: Accepted to WiML @ NeurIPS 2024 (extended version)

☆ How much can we forget about Data Contamination?

Sebastian Bordt, Suraj Srinivas, Valentyn Boreiko, Ulrike von Luxburg

The leakage of benchmark data into the training data has emerged as a significant challenge for evaluating the capabilities of large language models (LLMs). In this work, we use experimental evidence and theoretical estimates to challenge the common assumption that small-scale contamination renders benchmark evaluations invalid. First, we experimentally quantify the magnitude of benchmark overfitting based on scaling along three dimensions: The number of model parameters (up to 1.6B), the number of times an example is seen (up to 144), and the number of training tokens (up to 40B). We find that if model and data follow the Chinchilla scaling laws, minor contamination indeed leads to overfitting. At the same time, even 144 times of contamination can be forgotten if the training data is scaled beyond five times Chinchilla, a regime characteristic of many modern LLMs. We then derive a simple theory of example forgetting via cumulative weight decay. It allows us to bound the number of gradient steps required to forget past data for any training run where we know the hyperparameters of AdamW. This indicates that many LLMs, including Llama 3, have forgotten the data seen at the beginning of training. Experimentally, we demonstrate that forgetting occurs faster than what is predicted by our bounds. Taken together, our results suggest that moderate amounts of contamination can be forgotten at the end of realistically scaled training runs.

☆ Beyond Film Subtitles: Is YouTube the Best Approximation of Spoken Vocabulary? COLING 2025

Adam Nohejl, Frederikus Hudi, Eunike Andriani Kardinata, Shintaro Ozaki, Maria Angelica Riera Machin, Hongyu Sun, Justin Vasselli, Taro Watanabe

Word frequency is a key variable in psycholinguistics, useful for modeling human familiarity with words even in the era of large language models (LLMs). Frequency in film subtitles has proved to be a particularly good approximation of everyday language exposure. For many languages, however, film subtitles are not easily available, or are overwhelmingly translated from English. We demonstrate that frequencies extracted from carefully processed YouTube subtitles provide an approximation comparable to, and often better than, the best currently available resources. Moreover, they are available for languages for which a high-quality subtitle or speech corpus does not exist. We use YouTube subtitles to construct frequency norms for five diverse languages, Chinese, English, Indonesian, Japanese, and Spanish, and evaluate their correlation with lexical decision time, word familiarity, and lexical complexity. In addition to being strongly correlated with two psycholinguistic variables, a simple linear regression on the new frequencies achieves a new high score on a lexical complexity prediction task in English and Japanese, surpassing both models trained on film subtitle frequencies and the LLM GPT-4. Our code, the frequency lists, fastText word embeddings, and statistical language models are freely available at https://github.com/naist-nlp/tubelex.

comment: Submitted for review to COLING 2025. 8 pages, 3 figures

☆ Showing LLM-Generated Code Selectively Based on Confidence of LLMs

Jia Li, Yuqi Zhu, Yongmin Li, Ge Li, Zhi Jin

Large Language Models (LLMs) have shown impressive abilities in code generation, but they may generate erroneous programs. Reading a program takes ten times longer than writing it. Showing these erroneous programs to developers will waste developers' energies and introduce security risks to software. To address the above limitations, we propose HonestCoder, a novel LLM-based code generation approach. HonestCoder selectively shows the generated programs to developers based on LLMs' confidence. The confidence provides valuable insights into the correctness of generated programs. To achieve this goal, we propose a novel approach to estimate LLMs' confidence in code generation. It estimates confidence by measuring the multi-modal similarity between LLMs-generated programs. We collect and release a multilingual benchmark named TruthCodeBench, which consists of 2,265 samples and covers two popular programming languages (i.e., Python and Java). We apply HonestCoder to four popular LLMs (e.g., DeepSeek-Coder and Code Llama) and evaluate it on TruthCodeBench. Based on the experiments, we obtain the following insights. (1) HonestCoder can effectively estimate LLMs' confidence and accurately determine the correctness of generated programs. For example, HonestCoder outperforms the state-of-the-art baseline by 27.79% in AUROC and 63.74% in AUCPR. (2) HonestCoder can decrease the number of erroneous programs shown to developers. Compared to eight baselines, it can show more correct programs and fewer erroneous programs to developers. (3) Compared to showing code indiscriminately, HonestCoder only adds slight time overhead (approximately 0.4 seconds per requirement). (4) We discuss future directions to facilitate the application of LLMs in software development. We hope this work can motivate broad discussions about measuring the reliability of LLMs' outputs in performing code-related tasks.

☆ ALR$^2$: A Retrieve-then-Reason Framework for Long-context Question Answering

Huayang Li, Pat Verga, Priyanka Sen, Bowen Yang, Vijay Viswanathan, Patrick Lewis, Taro Watanabe, Yixuan Su

The context window of large language models (LLMs) has been extended significantly in recent years. However, while the context length that the LLM can process has grown, the capability of the model to accurately reason over that context degrades noticeably. This occurs because modern LLMs often become overwhelmed by the vast amount of information in the context; when answering questions, the model must identify and reason over relevant evidence sparsely distributed throughout the text. To alleviate the challenge of long-context reasoning, we develop a retrieve-then-reason framework, enabling LLMs to reason over relevant evidence collected during an intermediate retrieval step. We find that modern LLMs struggle to accurately retrieve relevant facts and instead, often hallucinate "retrieved facts", resulting in flawed reasoning and the production of incorrect answers. To address these issues, we introduce ALR$^2$, a method that augments the long-context reasoning capability of LLMs via an explicit two-stage procedure, i.e., aligning LLMs with the objectives of both retrieval and reasoning. We demonstrate the efficacy of ALR$^2$ for mitigating performance degradation in long-context reasoning tasks. Through extensive experiments on long-context QA benchmarks, we find our method to outperform competitive baselines by large margins, achieving at least 8.4 and 7.9 EM gains on the long-context versions of HotpotQA and SQuAD datasets, respectively.

☆ Consultation on Industrial Machine Faults with Large language Models

Apiradee Boonmee, Kritsada Wongsuwan, Pimchanok Sukjai

Industrial machine fault diagnosis is a critical component of operational efficiency and safety in manufacturing environments. Traditional methods rely heavily on expert knowledge and specific machine learning models, which can be limited in their adaptability and require extensive labeled data. This paper introduces a novel approach leveraging Large Language Models (LLMs), specifically through a structured multi-round prompting technique, to improve fault diagnosis accuracy. By dynamically crafting prompts, our method enhances the model's ability to synthesize information from diverse data sources, leading to improved contextual understanding and actionable recommendations. Experimental results demonstrate that our approach outperforms baseline models, achieving an accuracy of 91% in diagnosing various fault types. The findings underscore the potential of LLMs in revolutionizing industrial fault consultation practices, paving the way for more effective maintenance strategies in complex environments.

comment: 9 pages

☆ NLIP_Lab-IITH Low-Resource MT System for WMT24 Indic MT Shared Task

Pramit Sahoo, Maharaj Brahma, Maunendra Sankar Desarkar

In this paper, we describe our system for the WMT 24 shared task of Low-Resource Indic Language Translation. We consider eng $\leftrightarrow$ {as, kha, lus, mni} as participating language pairs. In this shared task, we explore the finetuning of a pre-trained model motivated by the pre-trained objective of aligning embeddings closer by alignment augmentation \cite{lin-etal-2020-pre} for 22 scheduled Indian languages. Our primary system is based on language-specific finetuning on a pre-trained model. We achieve chrF2 scores of 50.6, 42.3, 54.9, and 66.3 on the official public test set for eng$\rightarrow$as, eng$\rightarrow$kha, eng$\rightarrow$lus, eng$\rightarrow$mni respectively. We also explore multilingual training with/without language grouping and layer-freezing. Our code, models, and generated translations are available here: https://github.com/pramitsahoo/WMT2024-LRILT.

comment: WMT2024 INDICMT Shared Task

☆ Learning Semantic Structure through First-Order-Logic Translation EMNLP 2024

Akshay Chaturvedi, Nicholas Asher

In this paper, we study whether transformer-based language models can extract predicate argument structure from simple sentences. We firstly show that language models sometimes confuse which predicates apply to which objects. To mitigate this, we explore two tasks: question answering (Q/A), and first order logic (FOL) translation, and two regimes, prompting and finetuning. In FOL translation, we finetune several large language models on synthetic datasets designed to gauge their generalization abilities. For Q/A, we finetune encoder models like BERT and RoBERTa and use prompting for LLMs. The results show that FOL translation for LLMs is better suited to learn predicate argument structure.

comment: EMNLP 2024 Findings

☆ PersoBench: Benchmarking Personalized Response Generation in Large Language Models

Saleh Afzoon, Usman Naseem, Amin Beheshti, Zahra Jamali

While large language models (LLMs) have exhibited impressive conversational capabilities, their proficiency in delivering personalized responses remains unclear. Although recent benchmarks automatically evaluate persona consistency in role-playing contexts using LLM-based judgment, the evaluation of personalization in response generation remains underexplored. To address this gap, we present a new benchmark, PersoBench, to evaluate the personalization ability of LLMs in persona-aware dialogue generation within a zero-shot setting. We assess the performance of three open-source and three closed-source LLMs using well-known datasets and a range of metrics. Our analysis, conducted on three well-known persona-aware datasets, evaluates multiple dimensions of response quality, including fluency, diversity, coherence, and personalization, across both standard and chain-of-thought prompting methods. Our findings reveal that while LLMs excel at generating fluent and diverse responses, they are far from satisfactory in delivering personalized and coherent responses considering both the conversation context and the provided personas. Our benchmark implementation is available at https://github.com/salehafzoon/PersoBench.

☆ Cross-lingual Transfer for Automatic Question Generation by Learning Interrogative Structures in Target Languages EMNLP 2024

Seonjeong Hwang, Yunsu Kim, Gary Geunbae Lee

Automatic question generation (QG) serves a wide range of purposes, such as augmenting question-answering (QA) corpora, enhancing chatbot systems, and developing educational materials. Despite its importance, most existing datasets predominantly focus on English, resulting in a considerable gap in data availability for other languages. Cross-lingual transfer for QG (XLT-QG) addresses this limitation by allowing models trained on high-resource language datasets to generate questions in low-resource languages. In this paper, we propose a simple and efficient XLT-QG method that operates without the need for monolingual, parallel, or labeled data in the target language, utilizing a small language model. Our model, trained solely on English QA datasets, learns interrogative structures from a limited set of question exemplars, which are then applied to generate questions in the target language. Experimental results show that our method outperforms several XLT-QG baselines and achieves performance comparable to GPT-3.5-turbo across different languages. Additionally, the synthetic data generated by our model proves beneficial for training multilingual QA models. With significantly fewer parameters than large language models and without requiring additional training for target languages, our approach offers an effective solution for QG and QA tasks across various languages.

comment: EMNLP 2024

☆ Parallel Corpus Augmentation using Masked Language Models

Vibhuti Kumari, Narayana Murthy Kavi

In this paper we propose a novel method of augmenting parallel text corpora which promises good quality and is also capable of producing many fold larger corpora than the seed corpus we start with. We do not need any additional monolingual corpora. We use Multi-Lingual Masked Language Model to mask and predict alternative words in context and we use Sentence Embeddings to check and select sentence pairs which are likely to be translations of each other. We cross check our method using metrics for MT Quality Estimation. We believe this method can greatly alleviate the data scarcity problem for all language pairs for which a reasonable seed corpus is available.

comment: 21 Pages, 3 Figures. arXiv admin note: text overlap with arXiv:2011.01536 by other authors

☆ Generating bilingual example sentences with large language models as lexicography assistants

Raphael Merx, Ekaterina Vylomova, Kemal Kurniawan

We present a study of LLMs' performance in generating and rating example sentences for bilingual dictionaries across languages with varying resource levels: French (high-resource), Indonesian (mid-resource), and Tetun (low-resource), with English as the target language. We evaluate the quality of LLM-generated examples against the GDEX (Good Dictionary EXample) criteria: typicality, informativeness, and intelligibility. Our findings reveal that while LLMs can generate reasonably good dictionary examples, their performance degrades significantly for lower-resourced languages. We also observe high variability in human preferences for example quality, reflected in low inter-annotator agreement rates. To address this, we demonstrate that in-context learning can successfully align LLMs with individual annotator preferences. Additionally, we explore the use of pre-trained language models for automated rating of examples, finding that sentence perplexity serves as a good proxy for typicality and intelligibility in higher-resourced languages. Our study also contributes a novel dataset of 600 ratings for LLM-generated sentence pairs, and provides insights into the potential of LLMs in reducing the cost of lexicographic work, particularly for low-resource languages.

☆ Kiss up, Kick down: Exploring Behavioral Changes in Multi-modal Large Language Models with Assigned Visual Personas EMNLP 2024

Seungjong Sun, Eungu Lee, Seo Yeon Baek, Seunghyun Hwang, Wonbyung Lee, Dongyan Nan, Bernard J. Jansen, Jang Hyun Kim

This study is the first to explore whether multi-modal large language models (LLMs) can align their behaviors with visual personas, addressing a significant gap in the literature that predominantly focuses on text-based personas. We developed a novel dataset of 5K fictional avatar images for assignment as visual personas to LLMs, and analyzed their negotiation behaviors based on the visual traits depicted in these images, with a particular focus on aggressiveness. The results indicate that LLMs assess the aggressiveness of images in a manner similar to humans and output more aggressive negotiation behaviors when prompted with an aggressive visual persona. Interestingly, the LLM exhibited more aggressive negotiation behaviors when the opponent's image appeared less aggressive than their own, and less aggressive behaviors when the opponents image appeared more aggressive.

comment: EMNLP 2024

☆ Autoregressive Large Language Models are Computationally Universal

Dale Schuurmans, Hanjun Dai, Francesco Zanini

We show that autoregressive decoding of a transformer-based language model can realize universal computation, without external intervention or modification of the model's weights. Establishing this result requires understanding how a language model can process arbitrarily long inputs using a bounded context. For this purpose, we consider a generalization of autoregressive decoding where, given a long input, emitted tokens are appended to the end of the sequence as the context window advances. We first show that the resulting system corresponds to a classical model of computation, a Lag system, that has long been known to be computationally universal. By leveraging a new proof, we show that a universal Turing machine can be simulated by a Lag system with 2027 production rules. We then investigate whether an existing large language model can simulate the behaviour of such a universal Lag system. We give an affirmative answer by showing that a single system-prompt can be developed for gemini-1.5-pro-001 that drives the model, under deterministic (greedy) decoding, to correctly apply each of the 2027 production rules. We conclude that, by the Church-Turing thesis, prompted gemini-1.5-pro-001 with extended autoregressive (greedy) decoding is a general purpose computer.

comment: 32 pages

☆ Can Watermarked LLMs be Identified by Users via Crafted Prompts?

Aiwei Liu, Sheng Guan, Yiming Liu, Leyi Pan, Yifei Zhang, Liancheng Fang, Lijie Wen, Philip S. Yu, Xuming Hu

Text watermarking for Large Language Models (LLMs) has made significant progress in detecting LLM outputs and preventing misuse. Current watermarking techniques offer high detectability, minimal impact on text quality, and robustness to text editing. However, current researches lack investigation into the imperceptibility of watermarking techniques in LLM services. This is crucial as LLM providers may not want to disclose the presence of watermarks in real-world scenarios, as it could reduce user willingness to use the service and make watermarks more vulnerable to attacks. This work is the first to investigate the imperceptibility of watermarked LLMs. We design an identification algorithm called Water-Probe that detects watermarks through well-designed prompts to the LLM. Our key motivation is that current watermarked LLMs expose consistent biases under the same watermark key, resulting in similar differences across prompts under different watermark keys. Experiments show that almost all mainstream watermarking algorithms are easily identified with our well-designed prompts, while Water-Probe demonstrates a minimal false positive rate for non-watermarked LLMs. Finally, we propose that the key to enhancing the imperceptibility of watermarked LLMs is to increase the randomness of watermark key selection. Based on this, we introduce the Water-Bag strategy, which significantly improves watermark imperceptibility by merging multiple watermark keys.

comment: 25 pages, 5 figures, 8 tables

☆ Exploring Learnability in Memory-Augmented Recurrent Neural Networks: Precision, Stability, and Empirical Insights

Shrabon Das, Ankur Mali

This study explores the learnability of memory-less and memory-augmented RNNs, which are theoretically equivalent to Pushdown Automata. Empirical results show that these models often fail to generalize on longer sequences, relying more on precision than mastering symbolic grammar. Experiments on fully trained and component-frozen models reveal that freezing the memory component significantly improves performance, achieving state-of-the-art results on the Penn Treebank dataset (test perplexity reduced from 123.5 to 120.5). Models with frozen memory retained up to 90% of initial performance on longer sequences, compared to a 60% drop in standard models. Theoretical analysis suggests that freezing memory stabilizes temporal dependencies, leading to robust convergence. These findings stress the need for stable memory designs and long-sequence evaluations to understand RNNs true learnability limits.

comment: 21 pages, 4 theorems, 5 tables

☆ Media Framing through the Lens of Event-Centric Narratives EMNLP 2024

Rohan Das, Aditya Chandra, I-Ta Lee, Maria Leonor Pacheco

From a communications perspective, a frame defines the packaging of the language used in such a way as to encourage certain interpretations and to discourage others. For example, a news article can frame immigration as either a boost or a drain on the economy, and thus communicate very different interpretations of the same phenomenon. In this work, we argue that to explain framing devices we have to look at the way narratives are constructed. As a first step in this direction, we propose a framework that extracts events and their relations to other events, and groups them into high-level narratives that help explain frames in news articles. We show that our framework can be used to analyze framing in U.S. news for two different domains: immigration and gun control.

comment: Accepted to the 6th Workshop on Narrative Understanding, co-located with EMNLP 2024

☆ Analysis and Detection of Differences in Spoken User Behaviors between Autonomous and Wizard-of-Oz Systems

Mikey Elmers, Koji Inoue, Divesh Lala, Keiko Ochi, Tatsuya Kawahara

This study examined users' behavioral differences in a large corpus of Japanese human-robot interactions, comparing interactions between a tele-operated robot and an autonomous dialogue system. We analyzed user spoken behaviors in both attentive listening and job interview dialogue scenarios. Results revealed significant differences in metrics such as speech length, speaking rate, fillers, backchannels, disfluencies, and laughter between operator-controlled and autonomous conditions. Furthermore, we developed predictive models to distinguish between operator and autonomous system conditions. Our models demonstrated higher accuracy and precision compared to the baseline model, with several models also achieving a higher F1 score than the baseline.

comment: Accepted and will be presented at the 27th conference of the Oriental COCOSDA (O-COCOSDA 2024)

♻ ☆ Tokenization Falling Short: On Subword Robustness in Large Language Models EMNLP 2024

Yekun Chai, Yewei Fang, Qiwei Peng, Xuhong Li

Language models typically tokenize raw text into sequences of subword identifiers from a predefined vocabulary, a process inherently sensitive to typographical errors, length variations, and largely oblivious to the internal structure of tokens--issues we term the curse of tokenization. In this study, we delve into these drawbacks and demonstrate that large language models (LLMs) remain susceptible to these problems. This study systematically investigates these challenges and their impact on LLMs through three critical research questions: (1) complex problem solving, (2) token structure probing, and (3) resilience to typographical variation. Our findings reveal that scaling model parameters can mitigate the issue of tokenization; however, LLMs still suffer from biases induced by typos and other text format variations. Our experiments show that subword regularization such as BPE-dropout can mitigate this issue. We release our evaluation code and data at https://github.com/FloatAI/TKEval.

comment: EMNLP 2024 Findings

♻ ☆ Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge

Jiayi Ye, Yanbo Wang, Yue Huang, Dongping Chen, Qihui Zhang, Nuno Moniz, Tian Gao, Werner Geyer, Chao Huang, Pin-Yu Chen, Nitesh V Chawla, Xiangliang Zhang

LLM-as-a-Judge has been widely utilized as an evaluation method in various benchmarks and served as supervised rewards in model training. However, despite their excellence in many domains, potential issues are under-explored, undermining their reliability and the scope of their utility. Therefore, we identify 12 key potential biases and propose a new automated bias quantification framework-CALM-which systematically quantifies and analyzes each type of bias in LLM-as-a-Judge by using automated and principle-guided modification. Our experiments cover multiple popular language models, and the results indicate that while advanced models have achieved commendable overall performance, significant biases persist in certain specific tasks. Empirical results suggest that there remains room for improvement in the reliability of LLM-as-a-Judge. Moreover, we also discuss the explicit and implicit influence of these biases and give some suggestions for the reliable application of LLM-as-a-Judge. Our work highlights the need for stakeholders to address these issues and remind users to exercise caution in LLM-as-a-Judge applications.

♻ ☆ Video Instruction Tuning With Synthetic Data

Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, Chunyuan Li

The development of video large multimodal models (LMMs) has been hindered by the difficulty of curating large amounts of high-quality raw data from the web. To address this, we propose an alternative approach by creating a high-quality synthetic dataset specifically for video instruction-following, namely LLaVA-Video-178K. This dataset includes key tasks such as detailed captioning, open-ended question-answering (QA), and multiple-choice QA. By training on this dataset, in combination with existing visual instruction tuning data, we introduce LLaVA-Video, a new video LMM. Our experiments demonstrate that LLaVA-Video achieves strong performance across various video benchmarks, highlighting the effectiveness of our dataset. We plan to release the dataset, its generation pipeline, and the model checkpoints.

comment: Project page: https://llava-vl.github.io/blog/2024-09-30-llava-video/

♻ ☆ StorySparkQA: Expert-Annotated QA Pairs with Real-World Knowledge for Children's Story-Based Learning EMNLP 2024

Jiaju Chen, Yuxuan Lu, Shao Zhang, Bingsheng Yao, Yuanzhe Dong, Ying Xu, Yunyao Li, Qianwen Wang, Dakuo Wang, Yuling Sun

Interactive story reading is a common parent-child activity, where parents expect to teach both language skills and real-world knowledge beyond the story. While increasing storytelling and reading systems have been developed for this activity, they often fail to infuse real-world knowledge into the conversation. This limitation can be attributed to the existing question-answering (QA) datasets used for children's education, upon which the systems are built, failing to capture the nuances of how education experts think when conducting interactive story reading activities. To bridge this gap, we design an annotation framework, empowered by existing knowledge graph to capture experts' annotations and thinking process, and leverage this framework to construct StorySparkQA dataset, which comprises 5,868 expert-annotated QA pairs with real-world knowledge. We conduct automated and human expert evaluations across various QA pair generation settings to demonstrate that our StorySparkQA can effectively support models in generating QA pairs that target real-world knowledge beyond story content. StorySparkQA is available at https://huggingface.co/datasets/NEU-HAI/StorySparkQA.

comment: Accepted at EMNLP 2024 Main Conference

♻ ☆ EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage

Zeyi Liao, Lingbo Mo, Chejian Xu, Mintong Kang, Jiawei Zhang, Chaowei Xiao, Yuan Tian, Bo Li, Huan Sun

Generalist web agents have demonstrated remarkable potential in autonomously completing a wide range of tasks on real websites, significantly boosting human productivity. However, web tasks, such as booking flights, usually involve users' PII, which may be exposed to potential privacy risks if web agents accidentally interact with compromised websites, a scenario that remains largely unexplored in the literature. In this work, we narrow this gap by conducting the first study on the privacy risks of generalist web agents in adversarial environments. First, we present a realistic threat model for attacks on the website, where we consider two adversarial targets: stealing users' specific PII or the entire user request. Then, we propose a novel attack method, termed Environmental Injection Attack (EIA). EIA injects malicious content designed to adapt well to environments where the agents operate and our work instantiates EIA specifically for privacy scenarios in web environments. We collect 177 action steps that involve diverse PII categories on realistic websites from the Mind2Web, and conduct experiments using one of the most capable generalist web agent frameworks to date. The results demonstrate that EIA achieves up to 70% ASR in stealing specific PII and 16% ASR for full user request. Additionally, by accessing the stealthiness and experimenting with a defensive system prompt, we indicate that EIA is hard to detect and mitigate. Notably, attacks that are not well adapted for a webpage can be detected via human inspection, leading to our discussion about the trade-off between security and autonomy. However, extra attackers' efforts can make EIA seamlessly adapted, rendering such supervision ineffective. Thus, we further discuss the defenses at the pre- and post-deployment stages of the websites without relying on human supervision and call for more advanced defense strategies.

comment: 29 pages

♻ ☆ Convolutional Variational Autoencoders for Spectrogram Compression in Automatic Speech Recognition

Olga Iakovenko, Ivan Bondarenko

For many Automatic Speech Recognition (ASR) tasks audio features as spectrograms show better results than Mel-frequency Cepstral Coefficients (MFCC), but in practice they are hard to use due to a complex dimensionality of a feature space. The following paper presents an alternative approach towards generating compressed spectrogram representation, based on Convolutional Variational Autoencoders (VAE). A Convolutional VAE model was trained on a subsample of the LibriSpeech dataset to reconstruct short fragments of audio spectrograms (25 ms) from a 13-dimensional embedding. The trained model for a 40-dimensional (300 ms) embedding was used to generate features for corpus of spoken commands on the GoogleSpeechCommands dataset. Using the generated features an ASR system was built and compared to the model with MFCC features.

comment: Theory and Practice of Natural Computing 9th International Conference, TPNC 2020, Taoyuan, Taiwan, 2020, Proceedings 9

♻ ☆ Latte: Latent Attention for Linear Time Transformers

Rares Dolga, Lucas Maystre, Marius Cobzarenco, David Barber

The time complexity of the standard attention mechanism in transformers scales quadratically with sequence length. We propose a probabilistic framework for attention, enabling us to derive a novel low-rank linear re-parameterisation of both bidirectional and causal cases, based on defining a latent variable model. Our method can be seamlessly integrated as a drop-in replacement for the standard attention mechanism. Additionally, this framework provides a natural extension for combining local standard attention with our global linear attention. This approach allows us to extend the context length of existing large pre-trained models with only a few additional training steps. The resulting ``Latte Transformer'' achieves performance comparable to standard attention and other state-of-the-art models, while maintaining linear time and memory complexity, along with constant-time next-token prediction during inference.

♻ ☆ IoT-LLM: Enhancing Real-World IoT Task Reasoning with Large Language Models

Tuo An, Yunjiao Zhou, Han Zou, Jianfei Yang

Large Language Models (LLMs) have demonstrated remarkable capabilities across textual and visual domains but often generate outputs that violate physical laws, revealing a gap in their understanding of the physical world. Inspired by human cognition, where perception is fundamental to reasoning, we explore augmenting LLMs with enhanced perception abilities using Internet of Things (IoT) sensor data and pertinent knowledge for IoT task reasoning in the physical world. In this work, we systematically study LLMs capability to address real-world IoT tasks by augmenting their perception and knowledge base, and then propose a unified framework, IoT-LLM, to enhance such capability. In IoT-LLM, we customize three steps for LLMs: preprocessing IoT data into formats amenable to LLMs, activating their commonsense knowledge through chain-of-thought prompting and specialized role definitions, and expanding their understanding via IoT-oriented retrieval-augmented generation based on in-context learning. To evaluate the performance, We design a new benchmark with five real-world IoT tasks with different data types and reasoning difficulties and provide the benchmarking results on six open-source and close-source LLMs. Experimental results demonstrate the limitations of existing LLMs with naive textual inputs that cannot perform these tasks effectively. We show that IoT-LLM significantly enhances the performance of IoT tasks reasoning of LLM, such as GPT-4, achieving an average improvement of 65% across various tasks against previous methods. The results also showcase LLMs ability to comprehend IoT data and the physical law behind data by providing a reasoning process. Limitations of our work are claimed to inspire future research in this new era.

comment: 21 pages, 10 figures, under review

♻ ☆ How Much Can RAG Help the Reasoning of LLM?

Jingyu Liu, Jiaen Lin, Yong Liu

Retrieval-Augmented Generation (RAG) has gained significant popularity in modern Large Language Models (LLMs) due to its effectiveness in introducing new knowledge and reducing hallucinations. However, the deep understanding of RAG remains limited, how does RAG help the reasoning process and can RAG help improve the reasoning capability remains question. While external documents are typically considered as a method to incorporate domain-specific information, they also contain intermediate reasoning results related to the query, this suggests that documents could enhance the reasoning capability of LLMs, which has not been previously explored. In this paper, we investigate this issue in depth and find that while RAG can assist with reasoning, the help is limited. If we conceptualize the reasoning process as a tree with fixed depth, then RAG struggles to assist LLMs in performing deeper reasoning. Additionally, the information in the documents requires preprocessing to filter out noise. We demonstrate that this preprocessing is difficult to achieve simply fine-tuning of the LLM, it often necessitates numerous additional transformer layers to solve the problem. To simplify the problem, we propose DPrompt tuning, which effectively resolves the issue within just limited transformer layers, leading to improved performance.

♻ ☆ Annotation Guidelines for Corpus Novelties: Part 1 -- Named Entity Recognition

Arthur Amalvy, Vincent Labatut

The Novelties corpus is a collection of novels (and parts of novels) annotated for Named Entity Recognition (NER) among other tasks. This document describes the guidelines applied during its annotation. It contains the instructions used by the annotators, as well as a number of examples retrieved from the annotated novels, and illustrating expressions that should be marked as entities as well as expressions that should not.

♻ ☆ Retrieval-Augmented Hierarchical in-Context Reinforcement Learning and Hindsight Modular Reflections for Task Planning with LLMs

Chuanneng Sun, Songjun Huang, Dario Pompili

Large Language Models (LLMs) have demonstrated remarkable abilities in various language tasks, making them promising candidates for decision-making in robotics. Inspired by Hierarchical Reinforcement Learning (HRL), we propose Retrieval-Augmented in-context reinforcement Learning (RAHL), a novel framework that decomposes complex tasks into sub-tasks using an LLM-based high-level policy, in which a complex task is decomposed into sub-tasks by a high-level policy on-the-fly. The sub-tasks, defined by goals, are assigned to the low-level policy to complete. To improve the agent's performance in multi-episode execution, we propose Hindsight Modular Reflection (HMR), where, instead of reflecting on the full trajectory, we let the agent reflect on shorter sub-trajectories to improve reflection efficiency. We evaluated the decision-making ability of the proposed RAHL in three benchmark environments--ALFWorld, Webshop, and HotpotQA. The results show that RAHL can achieve an improvement in performance in 9%, 42%, and 10% in 5 episodes of execution in strong baselines. Furthermore, we also implemented RAHL on the Boston Dynamics SPOT robot. The experiment shows that the robot can scan the environment, find entrances, and navigate to new rooms controlled by the LLM policy.

♻ ☆ SaySelf: Teaching LLMs to Express Confidence with Self-Reflective Rationales EMNLP 2024

Tianyang Xu, Shujin Wu, Shizhe Diao, Xiaoze Liu, Xingyao Wang, Yangyi Chen, Jing Gao

Large language models (LLMs) often generate inaccurate or fabricated information and generally fail to indicate their confidence, which limits their broader applications. Previous work elicits confidence from LLMs by direct or self-consistency prompting, or constructing specific datasets for supervised finetuning. The prompting-based approaches have inferior performance, and the training-based approaches are limited to binary or inaccurate group-level confidence estimates. In this work, we present the advanced SaySelf, a training framework that teaches LLMs to express more accurate fine-grained confidence estimates. In addition, beyond the confidence scores, SaySelf initiates the process of directing LLMs to produce self-reflective rationales that clearly identify gaps in their parametric knowledge and explain their uncertainty. This is achieved by using an LLM to automatically summarize the uncertainties in specific knowledge via natural language. The summarization is based on the analysis of the inconsistency in multiple sampled reasoning chains, and the resulting data is utilized for supervised fine-tuning. Moreover, we utilize reinforcement learning with a meticulously crafted reward function to calibrate the confidence estimates, motivating LLMs to deliver accurate, high-confidence predictions and to penalize overconfidence in erroneous outputs. Experimental results in both in-distribution and out-of-distribution datasets demonstrate the effectiveness of SaySelf in reducing the confidence calibration error and maintaining the task performance. We show that the generated self-reflective rationales are reasonable and can further contribute to the calibration. The code is made public at https://github.com/xu1868/SaySelf.

comment: EMNLP 2024 Main

♻ ☆ Language Model Empowered Spatio-Temporal Forecasting via Physics-Aware Reprogramming

Hao Wang, Jindong Han, Wei Fan, Hao Liu

Spatio-temporal forecasting is pivotal in numerous real-world applications, including transportation planning, energy management, and climate monitoring. In this work, we aim to harness the reasoning and generalization abilities of Pre-trained Language Models (PLMs) for more effective spatio-temporal forecasting, particularly in data-scarce scenarios. However, recent studies uncover that PLMs, which are primarily trained on textual data, often falter when tasked with modeling the intricate correlations in numerical time series, thereby limiting their effectiveness in comprehending spatio-temporal data. To bridge the gap, we propose RePST, a physics-aware PLM reprogramming framework tailored for spatio-temporal forecasting. Specifically, we first propose a physics-aware decomposer that adaptively disentangles spatially correlated time series into interpretable sub-components, which facilitates PLM to understand sophisticated spatio-temporal dynamics via a divide-and-conquer strategy. Moreover, we propose a selective discrete reprogramming scheme, which introduces an expanded spatio-temporal vocabulary space to project spatio-temporal series into discrete representations. This scheme minimizes the information loss during reprogramming and enriches the representations derived by PLMs. Extensive experiments on real-world datasets show that the proposed RePST outperforms twelve state-of-the-art baseline methods, particularly in data-scarce scenarios, highlighting the effectiveness and superior generalization capabilities of PLMs for spatio-temporal forecasting.

♻ ☆ Improving Statistical Significance in Human Evaluation of Automatic Metrics via Soft Pairwise Accuracy

Brian Thompson, Nitika Mathur, Daniel Deutsch, Huda Khayrallah

Selecting an automatic metric that best emulates human annotators is often non-trivial, because there is no clear definition of "best emulates." A meta-metric is required to compare the human judgments to the automatic metric scores, and metric rankings depend on the choice of meta-metric. We propose Soft Pairwise Accuracy (SPA), a new meta-metric that builds on Pairwise Accuracy (PA) but incorporates the statistical significance of both the human judgments and the metric scores. We show that SPA is more stable than PA with respect to changes in the number of systems/segments used for evaluation. We also show that PA can only assign a small set of distinct output values to metrics, and this results in many metrics being artificially assigned the exact same PA score. We demonstrate that SPA fixes this issue. Finally, we show that SPA is more discriminative than PA, producing more statistically significant comparisons between metrics. SPA was selected as the official system-level metric for the 2024 WMT Metrics Shared Task.

comment: Accepted at WMT 2024

♻ ☆ MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making

Yubin Kim, Chanwoo Park, Hyewon Jeong, Yik Siu Chan, Xuhai Xu, Daniel McDuff, Hyeonhoon Lee, Marzyeh Ghassemi, Cynthia Breazeal, Hae Won Park

Foundation models are becoming valuable tools in medicine. Yet despite their promise, the best way to leverage Large Language Models (LLMs) in complex medical tasks remains an open question. We introduce a novel multi-agent framework, named Medical Decision-making Agents (MDAgents) that helps address this gap by automatically assigning a collaboration structure to a team of LLMs. The assigned solo or group collaboration structure is tailored to the medical task at hand, emulating real-world medical decision-making processes adapted to tasks of varying complexities. We evaluate our framework and baseline methods using state-of-the-art LLMs across a suite of real-world medical knowledge and medical diagnosis benchmarks. MDAgents achieved the best performance in seven out of ten benchmarks on tasks requiring an understanding of medical knowledge and multi-modal reasoning, showing a significant improvement of up to 6.5% (p < 0.05) compared to previous methods' best performances. Ablation studies reveal that MDAgents effectively determines medical complexity to optimize for efficiency and accuracy across diverse medical tasks. Notably, the combination of moderator review and external medical knowledge in group collaboration resulted in an average accuracy improvement of 11.8%. Our code can be found at https://github.com/mitmedialab/MDAgents.

♻ ☆ Unraveling the Truth: Do VLMs really Understand Charts? A Deep Dive into Consistency and Robustness

Srija Mukhopadhyay, Adnan Qidwai, Aparna Garimella, Pritika Ramu, Vivek Gupta, Dan Roth

Chart question answering (CQA) is a crucial area of Visual Language Understanding. However, the robustness and consistency of current Visual Language Models (VLMs) in this field remain under-explored. This paper evaluates state-of-the-art VLMs on comprehensive datasets, developed specifically for this study, encompassing diverse question categories and chart formats. We investigate two key aspects: 1) the models' ability to handle varying levels of chart and question complexity, and 2) their robustness across different visual representations of the same underlying data. Our analysis reveals significant performance variations based on question and chart types, highlighting both strengths and weaknesses of current models. Additionally, we identify areas for improvement and propose future research directions to build more robust and reliable CQA systems. This study sheds light on the limitations of current models and paves the way for future advancements in the field.

comment: 22 pages, 9 Tables, 5 figures, 22 examples

♻ ☆ Unlocking Anticipatory Text Generation: A Constrained Approach for Large Language Models Decoding EMNLP 2024

Lifu Tu, Semih Yavuz, Jin Qu, Jiacheng Xu, Rui Meng, Caiming Xiong, Yingbo Zhou

Large Language Models (LLMs) have demonstrated a powerful ability for text generation. However, achieving optimal results with a given prompt or instruction can be challenging, especially for billion-sized models. Additionally, undesired behaviors such as toxicity or hallucinations can manifest. While much larger models (e.g., ChatGPT) may demonstrate strength in mitigating these issues, there is still no guarantee of complete prevention. In this work, we propose formalizing text generation as a future-constrained generation problem to minimize undesirable behaviors and enforce faithfulness to instructions. The estimation of future constraint satisfaction, accomplished using LLMs, guides the text generation process. Our extensive experiments demonstrate the effectiveness of the proposed approach across three distinct text generation tasks: keyword-constrained generation (Lin et al., 2020), toxicity reduction (Gehman et al., 2020), and factual correctness in question-answering (Gao et al., 2023).

comment: EMNLP 2024 Main

♻ ☆ Why Would You Suggest That? Human Trust in Language Model Responses

Manasi Sharma, Ho Chit Siu, Rohan Paleja, Jaime D. Peña

The emergence of Large Language Models (LLMs) has revealed a growing need for human-AI collaboration, especially in creative decision-making scenarios where trust and reliance are paramount. Through human studies and model evaluations on the open-ended News Headline Generation task from the LaMP benchmark, we analyze how the framing and presence of explanations affect user trust and model performance. Overall, we provide evidence that adding an explanation in the model response to justify its reasoning significantly increases self-reported user trust in the model when the user has the opportunity to compare various responses. Position and faithfulness of these explanations are also important factors. However, these gains disappear when users are shown responses independently, suggesting that humans trust all model responses, including deceptive ones, equitably when they are shown in isolation. Our findings urge future research to delve deeper into the nuanced evaluation of trust in human-machine teaming systems.

♻ ☆ Evaluating $n$-Gram Novelty of Language Models Using Rusty-DAWG EMNLP 2024

William Merrill, Noah A. Smith, Yanai Elazar

How novel are texts generated by language models (LMs) relative to their training corpora? In this work, we investigate the extent to which modern LMs generate $n$-grams from their training data, evaluating both (i) the probability LMs assign to complete training $n$-grams and (ii) $n$-novelty, the proportion of $n$-grams generated by an LM that did not appear in the training data (for arbitrarily large $n$). To enable arbitrary-length $n$-gram search over a corpus in constant time w.r.t. corpus size, we develop Rusty-DAWG, a novel search tool inspired by indexing of genomic data. We compare the novelty of LM-generated text to human-written text and explore factors that affect generation novelty, focusing on the Pythia models. We find that, for $n > 4$, LM-generated text is less novel than human-written text, though it is more novel for smaller $n$. Larger LMs and more constrained decoding strategies both decrease novelty. Finally, we show that LMs complete $n$-grams with lower loss if they are more frequent in the training data. Overall, our results reveal factors influencing the novelty of LM-generated text, and we release Rusty-DAWG to facilitate further pretraining data research.

comment: To appear at EMNLP 2024

♻ ☆ Probabilities of Chat LLMs Are Miscalibrated but Still Predict Correctness on Multiple-Choice Q&A

Benjamin Plaut, Nguyen X. Khanh, Tu Trinh

We study 14 large language models (LLMs) fine-tuned for chat and find that their maximum softmax probabilities (MSPs) are consistently miscalibrated on multiple-choice Q&A. However, those MSPs might still encode useful uncertainty information. Specifically, we hypothesized that wrong answers would be associated with smaller MSPs compared to correct answers. Via rigororous statistical testing, we show that this hypothesis holds for models which perform well on the underlying Q&A task. We also find a strong direction correlation between Q&A accuracy and MSP correctness prediction, while finding no correlation between Q&A accuracy and calibration error. This suggests that within the current fine-tuning paradigm, we can expect correctness prediction but not calibration to improve as LLM capabilities progress. To demonstrate the utility of correctness prediction, we show that when models have the option to abstain, performance can be improved by selectively abstaining based on the MSP of the initial model response, using only a small amount of labeled data to choose the MSP threshold.

♻ ☆ MetaKP: On-Demand Keyphrase Generation EMNLP 2024

Di Wu, Xiaoxian Shen, Kai-Wei Chang

Traditional keyphrase prediction methods predict a single set of keyphrases per document, failing to cater to the diverse needs of users and downstream applications. To bridge the gap, we introduce on-demand keyphrase generation, a novel paradigm that requires keyphrases that conform to specific high-level goals or intents. For this task, we present MetaKP, a large-scale benchmark comprising four datasets, 7500 documents, and 3760 goals across news and biomedical domains with human-annotated keyphrases. Leveraging MetaKP, we design both supervised and unsupervised methods, including a multi-task fine-tuning approach and a self-consistency prompting method with large language models. The results highlight the challenges of supervised fine-tuning, whose performance is not robust to distribution shifts. By contrast, the proposed self-consistency prompting approach greatly improves the performance of large language models, enabling GPT-4o to achieve 0.548 SemF1, surpassing the performance of a fully fine-tuned BART-base model. Finally, we demonstrate the potential of our method to serve as a general NLP infrastructure, exemplified by its application in epidemic event detection from social media.

comment: EMNLP 2024 (Findings)

♻ ☆ VideoCLIP-XL: Advancing Long Description Understanding for Video CLIP Models EMNLP 2024

Jiapeng Wang, Chengyu Wang, Kunzhe Huang, Jun Huang, Lianwen Jin

Contrastive Language-Image Pre-training (CLIP) has been widely studied and applied in numerous applications. However, the emphasis on brief summary texts during pre-training prevents CLIP from understanding long descriptions. This issue is particularly acute regarding videos given that videos often contain abundant detailed contents. In this paper, we propose the VideoCLIP-XL (eXtra Length) model, which aims to unleash the long-description understanding capability of video CLIP models. Firstly, we establish an automatic data collection system and gather a large-scale VILD pre-training dataset with VIdeo and Long-Description pairs. Then, we propose Text-similarity-guided Primary Component Matching (TPCM) to better learn the distribution of feature space while expanding the long description capability. We also introduce two new tasks namely Detail-aware Description Ranking (DDR) and Hallucination-aware Description Ranking (HDR) for further understanding improvement. Finally, we construct a Long Video Description Ranking (LVDR) benchmark for evaluating the long-description capability more comprehensively. Extensive experimental results on widely-used text-video retrieval benchmarks with both short and long descriptions and our LVDR benchmark can fully demonstrate the effectiveness of our method.

comment: EMNLP 2024 Main conference

♻ ☆ Identifying Factual Inconsistencies in Summaries: Grounding LLM Inference via Task Taxonomy EMNLP 2024

Liyan Xu, Zhenlin Su, Mo Yu, Jin Xu, Jinho D. Choi, Jie Zhou, Fei Liu

Factual inconsistencies pose a significant hurdle for the faithful summarization by generative models. While a major direction to enhance inconsistency detection is to derive stronger Natural Language Inference (NLI) models, we propose an orthogonal aspect that underscores the importance of incorporating task-specific taxonomy into the inference. To this end, we consolidate key error types of inconsistent facts in summaries, and incorporate them to facilitate both the zero-shot and supervised paradigms of LLMs. Extensive experiments on ten datasets of five distinct domains suggest that, zero-shot LLM inference could benefit from the explicit solution space depicted by the error type taxonomy, and achieves state-of-the-art performance overall, surpassing specialized non-LLM baselines, as well as recent LLM baselines. We further distill models that fuse the taxonomy into parameters through our designed prompt completions and supervised training strategies, efficiently substituting state-of-the-art zero-shot inference with much larger LLMs.

comment: Accepted to EMNLP 2024 Findings

♻ ☆ To Err Is Human, but Llamas Can Learn It Too

Agnes Luhtaru, Taido Purason, Martin Vainikko, Maksym Del, Mark Fishel

This study explores enhancing grammatical error correction (GEC) through artificial error generation (AEG) using language models (LMs). Specifically, we fine-tune Llama 2-based LMs for error generation and find that this approach yields synthetic errors akin to human errors. Next, we train GEC Llama models with the help of these artificial errors and outperform previous state-of-the-art error correction models, with gains ranging between 0.8 and 6 F0.5 points across all tested languages (German, Ukrainian, and Estonian). Moreover, we demonstrate that generating errors by fine-tuning smaller sequence-to-sequence models and prompting large commercial LMs (GPT-3.5 and GPT-4) also results in synthetic errors beneficially affecting error generation models.

♻ ☆ "Seeing the Big through the Small": Can LLMs Approximate Human Judgment Distributions on NLI from a Few Explanations? EMNLP 2024

Beiduo Chen, Xinpeng Wang, Siyao Peng, Robert Litschko, Anna Korhonen, Barbara Plank

Human label variation (HLV) is a valuable source of information that arises when multiple human annotators provide different labels for valid reasons. In Natural Language Inference (NLI) earlier approaches to capturing HLV involve either collecting annotations from many crowd workers to represent human judgment distribution (HJD) or use expert linguists to provide detailed explanations for their chosen labels. While the former method provides denser HJD information, obtaining it is resource-intensive. In contrast, the latter offers richer textual information but it is challenging to scale up to many human judges. Besides, large language models (LLMs) are increasingly used as evaluators ("LLM judges") but with mixed results, and few works aim to study HJDs. This study proposes to exploit LLMs to approximate HJDs using a small number of expert labels and explanations. Our experiments show that a few explanations significantly improve LLMs' ability to approximate HJDs with and without explicit labels, thereby providing a solution to scale up annotations for HJD. However, fine-tuning smaller soft-label aware models with the LLM-generated model judgment distributions (MJDs) presents partially inconsistent results: while similar in distance, their resulting fine-tuned models and visualized distributions differ substantially. We show the importance of complementing instance-level distance measures with a global-level shape metric and visualization to more effectively evaluate MJDs against human judgment distributions.

comment: Accepted by EMNLP 2024 Findings, 24 pages, 9 figures

♻ ☆ A SMART Mnemonic Sounds like "Glue Tonic": Mixing LLMs with Student Feedback to Make Mnemonic Learning Stick EMNLP 2024

Nishant Balepur, Matthew Shu, Alexander Hoyle, Alison Robey, Shi Feng, Seraphina Goldfarb-Tarrant, Jordan Boyd-Graber

Keyword mnemonics are memorable explanations that link new terms to simpler keywords. Prior work generates mnemonics for students, but they do not train models using mnemonics students prefer and aid learning. We build SMART, a mnemonic generator trained on feedback from real students learning new terms. To train SMART, we first fine-tune LLaMA-2 on a curated set of user-written mnemonics. We then use LLM alignment to enhance SMART: we deploy mnemonics generated by SMART in a flashcard app to find preferences on mnemonics students favor. We gather 2684 preferences from 45 students across two types: expressed (inferred from ratings) and observed (inferred from student learning), yielding three key findings. First, expressed and observed preferences disagree; what students think is helpful does not always capture what is truly helpful. Second, Bayesian models can synthesize complementary data from multiple preference types into a single effectiveness signal. SMART is tuned via Direct Preference Optimization on this signal, which resolves ties and missing labels in the typical method of pairwise comparisons, augmenting data for LLM output quality gains. Third, mnemonic experts assess SMART as matching GPT-4 at much lower deployment costs, showing the utility of capturing diverse student feedback to align LLMs in education.

comment: EMNLP 2024

♻ ☆ Flow of Reasoning:Training LLMs for Divergent Problem Solving with Minimal Examples

Fangxu Yu, Lai Jiang, Haoqiang Kang, Shibo Hao, Lianhui Qin

The ability to generate diverse solutions to a given problem is a hallmark of human creativity. This divergent reasoning is also crucial for machines, enhancing their robustness and enabling them to assist humans in many applications such as scientific discovery. However, existing approaches to multi-step reasoning with large language models (LLMs) have mostly focused only on reasoning accuracy, without further discovering more diverse valid solutions. For example, supervised fine-tuning can improve LLM reasoning quality, but requires extensive supervised data to capture the full range of possible solutions. Reinforcement learning aims to find limited highest-reward solutions while neglecting the solution diversity. To fill this gap, we propose Flow of Reasoning (FoR), an efficient diversity-seeking LLM finetuning method aimed at improving reasoning quality and diversity with minimal data. FoR formulates multi-step LLM reasoning as a Markovian flow on a DAG-structured reasoning graph. This formulation allows us to incorporate and adapt principled GFlowNet approaches, for finetuning LLMs to sample diverse reasoning paths with probabilities proportional to the (unnormalized) reward of target problems. Extensive experiments show that, with limited training examples (e.g., 15 examples), FoR enables the discovery of diverse, creative, high-quality solutions, greatly outperforming a wide range of existing inference and training methods across five challenging puzzle-solving tasks, including BlocksWorld (embodied reasoning), Game24 (math puzzle solving), Rubik's Cube (spatial reasoning), 1D-ARC (abstraction reasoning), and PrOntoQA (logical reasoning). Code is available at https://github.com/Yu-Fangxu/FoR.

♻ ☆ Jailbreaking as a Reward Misspecification Problem

Zhihui Xie, Jiahui Gao, Lei Li, Zhenguo Li, Qi Liu, Lingpeng Kong

The widespread adoption of large language models (LLMs) has raised concerns about their safety and reliability, particularly regarding their vulnerability to adversarial attacks. In this paper, we propose a novel perspective that attributes this vulnerability to reward misspecification during the alignment process. This misspecification occurs when the reward function fails to accurately capture the intended behavior, leading to misaligned model outputs. We introduce a metric ReGap to quantify the extent of reward misspecification and demonstrate its effectiveness and robustness in detecting harmful backdoor prompts. Building upon these insights, we present ReMiss, a system for automated red teaming that generates adversarial prompts in a reward-misspecified space. ReMiss achieves state-of-the-art attack success rates on the AdvBench benchmark against various target aligned LLMs while preserving the human readability of the generated prompts. Furthermore, these attacks on open-source models demonstrate high transferability to closed-source models like GPT-4o and out-of-distribution tasks from HarmBench. Detailed analysis highlights the unique advantages of the proposed reward misspecification objective compared to previous methods, offering new insights for improving LLM safety and robustness.

♻ ☆ KARL: Knowledge-Aware Retrieval and Representations aid Retention and Learning in Students EMNLP 2024

Matthew Shu, Nishant Balepur, Shi Feng, Jordan Boyd-Graber

Flashcard schedulers rely on 1) student models to predict the flashcards a student knows; and 2) teaching policies to pick which cards to show next via these predictions. Prior student models, however, just use study data like the student's past responses, ignoring the text on cards. We propose content-aware scheduling, the first schedulers exploiting flashcard content. To give the first evidence that such schedulers enhance student learning, we build KARL, a simple but effective content-aware student model employing deep knowledge tracing (DKT), retrieval, and BERT to predict student recall. We train KARL by collecting a new dataset of 123,143 study logs on diverse trivia questions. KARL bests existing student models in AUC and calibration error. To ensure our improved predictions lead to better student learning, we create a novel delta-based teaching policy to deploy KARL online. Based on 32 study paths from 27 users, KARL improves learning efficiency over SOTA, showing KARL's strength and encouraging researchers to look beyond historical study data to fully capture student abilities.

comment: EMNLP 2024

♻ ☆ OpenHands: An Open Platform for AI Software Developers as Generalist Agents

Xingyao Wang, Boxuan Li, Yufan Song, Frank F. Xu, Xiangru Tang, Mingchen Zhuge, Jiayi Pan, Yueqi Song, Bowen Li, Jaskirat Singh, Hoang H. Tran, Fuqiang Li, Ren Ma, Mingzhang Zheng, Bill Qian, Yanjun Shao, Niklas Muennighoff, Yizhe Zhang, Binyuan Hui, Junyang Lin, Robert Brennan, Hao Peng, Heng Ji, Graham Neubig

Software is one of the most powerful tools that we humans have at our disposal; it allows a skilled programmer to interact with the world in complex and profound ways. At the same time, thanks to improvements in large language models (LLMs), there has also been a rapid development in AI agents that interact with and affect change in their surrounding environments. In this paper, we introduce OpenHands (f.k.a. OpenDevin), a platform for the development of powerful and flexible AI agents that interact with the world in similar ways to those of a human developer: by writing code, interacting with a command line, and browsing the web. We describe how the platform allows for the implementation of new agents, safe interaction with sandboxed environments for code execution, coordination between multiple agents, and incorporation of evaluation benchmarks. Based on our currently incorporated benchmarks, we perform an evaluation of agents over 15 challenging tasks, including software engineering (e.g., SWE-BENCH) and web browsing (e.g., WEBARENA), among others. Released under the permissive MIT license, OpenHands is a community project spanning academia and industry with more than 2.1K contributions from over 188 contributors.

comment: Code: https://github.com/All-Hands-AI/OpenHands

♻ ☆ To Know or Not To Know? Analyzing Self-Consistency of Large Language Models under Ambiguity EMNLP 2024

Anastasiia Sedova, Robert Litschko, Diego Frassinelli, Benjamin Roth, Barbara Plank

One of the major aspects contributing to the striking performance of large language models (LLMs) is the vast amount of factual knowledge accumulated during pre-training. Yet, many LLMs suffer from self-inconsistency, which raises doubts about their trustworthiness and reliability. This paper focuses on entity type ambiguity, analyzing the proficiency and consistency of state-of-the-art LLMs in applying factual knowledge when prompted with ambiguous entities. To do so, we propose an evaluation protocol that disentangles knowing from applying knowledge, and test state-of-the-art LLMs on 49 ambiguous entities. Our experiments reveal that LLMs struggle with choosing the correct entity reading, achieving an average accuracy of only 85%, and as low as 75% with underspecified prompts. The results also reveal systematic discrepancies in LLM behavior, showing that while the models may possess knowledge, they struggle to apply it consistently, exhibit biases toward preferred readings, and display self-inconsistencies. This highlights the need to address entity ambiguity in the future for more trustworthy LLMs.

comment: EMNLP 2024 Findings

♻ ☆ Soda-Eval: Open-Domain Dialogue Evaluation in the age of LLMs EMNLP2024

John Mendonça, Isabel Trancoso, Alon Lavie

Although human evaluation remains the gold standard for open-domain dialogue evaluation, the growing popularity of automated evaluation using Large Language Models (LLMs) has also extended to dialogue. However, most frameworks leverage benchmarks that assess older chatbots on aspects such as fluency and relevance, which are not reflective of the challenges associated with contemporary models. In fact, a qualitative analysis on Soda, a GPT-3.5 generated dialogue dataset, suggests that current chatbots may exhibit several recurring issues related to coherence and commonsense knowledge, but generally produce highly fluent and relevant responses. Noting the aforementioned limitations, this paper introduces Soda-Eval, an annotated dataset based on Soda that covers over 120K turn-level assessments across 10K dialogues, where the annotations were generated by GPT-4. Using Soda-Eval as a benchmark, we then study the performance of several open-access instruction-tuned LLMs, finding that dialogue evaluation remains challenging. Fine-tuning these models improves performance over few-shot inferences, both in terms of correlation and explanation.

comment: Accepted to EMNLP2024 (findings)

♻ ☆ MultiContrievers: Analysis of Dense Retrieval Representations

Seraphina Goldfarb-Tarrant, Pedro Rodriguez, Jane Dwivedi-Yu, Patrick Lewis

Dense retrievers compress source documents into (possibly lossy) vector representations, yet there is little analysis of what information is lost versus preserved, and how it affects downstream tasks. We conduct the first analysis of the information captured by dense retrievers compared to the language models they are based on (e.g., BERT versus Contriever). We use 25 MultiBert checkpoints as randomized initialisations to train MultiContrievers, a set of 25 contriever models. We test whether specific pieces of information -- such as gender and occupation -- can be extracted from contriever vectors of wikipedia-like documents. We measure this extractability via information theoretic probing. We then examine the relationship of extractability to performance and gender bias, as well as the sensitivity of these results to many random initialisations and data shuffles. We find that (1) contriever models have significantly increased extractability, but extractability usually correlates poorly with benchmark performance 2) gender bias is present, but is not caused by the contriever representations 3) there is high sensitivity to both random initialisation and to data shuffle, suggesting that future retrieval research should test across a wider spread of both.

♻ ☆ Logistic Regression makes small LLMs strong and explainable "tens-of-shot" classifiers

Marcus Buckmann, Edward Hill

For simple classification tasks, we show that users can benefit from the advantages of using small, local, generative language models instead of large commercial models without a trade-off in performance or introducing extra labelling costs. These advantages, including those around privacy, availability, cost, and explainability, are important both in commercial applications and in the broader democratisation of AI. Through experiments on 17 sentence classification tasks (2-4 classes), we show that penalised logistic regression on the embeddings from a small LLM equals (and usually betters) the performance of a large LLM in the "tens-of-shot" regime. This requires no more labelled instances than are needed to validate the performance of the large LLM. Finally, we extract stable and sensible explanations for classification decisions.

comment: 48 pages, 24 figures

♻ ☆ SpeciaLex: A Benchmark for In-Context Specialized Lexicon Learning EMNLP 2024

Joseph Marvin Imperial, Harish Tayyar Madabushi

Specialized lexicons are collections of words with associated constraints such as special definitions, specific roles, and intended target audiences. These constraints are necessary for content generation and documentation tasks (e.g., writing technical manuals or children's reading materials), where the goal is to reduce the ambiguity of text content and increase its overall readability for a specific group of audience. Understanding how large language models can capture these constraints can help researchers build better, more impactful tools for wider use beyond the NLP community. Towards this end, we introduce SpeciaLex, a benchmark for evaluating a language model's ability to follow specialized lexicon-based constraints across 18 diverse subtasks with 1,785 test instances covering core tasks of Checking, Identification, Rewriting, and Open Generation. We present an empirical evaluation of 15 open and closed-source LLMs and discuss insights on how factors such as model scale, openness, setup, and recency affect performance upon evaluating with the benchmark.

comment: Camera-ready for EMNLP 2024 (Findings)

♻ ☆ A LLM-Based Ranking Method for the Evaluation of Automatic Counter-Narrative Generation

Irune Zubiaga, Aitor Soroa, Rodrigo Agerri

This paper proposes a novel approach to evaluate Counter Narrative (CN) generation using a Large Language Model (LLM) as an evaluator. We show that traditional automatic metrics correlate poorly with human judgements and fail to capture the nuanced relationship between generated CNs and human perception. To alleviate this, we introduce a model ranking pipeline based on pairwise comparisons of generated CNs from different models, organized in a tournament-style format. The proposed evaluation method achieves a high correlation with human preference, with a $\rho$ score of 0.88. As an additional contribution, we leverage LLMs as zero-shot CN generators and provide a comparative analysis of chat, instruct, and base models, exploring their respective strengths and limitations. Through meticulous evaluation, including fine-tuning experiments, we elucidate the differences in performance and responsiveness to domain-specific data. We conclude that chat-aligned models in zero-shot are the best option for carrying out the task, provided they do not refuse to generate an answer due to security concerns.

♻ ☆ SLANG: New Concept Comprehension of Large Language Models

Lingrui Mei, Shenghua Liu, Yiwei Wang, Baolong Bi, Xueqi Cheng

The dynamic nature of language, particularly evident in the realm of slang and memes on the Internet, poses serious challenges to the adaptability of large language models (LLMs). Traditionally anchored to static datasets, these models often struggle to keep up with the rapid linguistic evolution characteristic of online communities. This research aims to bridge this gap by enhancing LLMs' comprehension of the evolving new concepts on the Internet, without the high cost of continual retraining. In pursuit of this goal, we introduce $\textbf{SLANG}$, a benchmark designed to autonomously integrate novel data and assess LLMs' ability to comprehend emerging concepts, alongside $\textbf{FOCUS}$, an approach uses causal inference to enhance LLMs to understand new phrases and their colloquial context. Our benchmark and approach involves understanding real-world instances of linguistic shifts, serving as contextual beacons, to form more precise and contextually relevant connections between newly emerging expressions and their meanings. The empirical analysis shows that our causal inference-based approach outperforms the baseline methods in terms of precision and relevance in the comprehension of Internet slang and memes.

♻ ☆ Self-Training for Sample-Efficient Active Learning for Text Classification with Pre-Trained Language Models EMNLP 2024

Christopher Schröder, Gerhard Heyer

Active learning is an iterative labeling process that is used to obtain a small labeled subset, despite the absence of labeled data, thereby enabling to train a model for supervised tasks such as text classification. While active learning has made considerable progress in recent years due to improvements provided by pre-trained language models, there is untapped potential in the often neglected unlabeled portion of the data, although it is available in considerably larger quantities than the usually small set of labeled data. In this work, we investigate how self-training, a semi-supervised approach that uses a model to obtain pseudo-labels for unlabeled data, can be used to improve the efficiency of active learning for text classification. Building on a comprehensive reproduction of four previous self-training approaches, some of which are evaluated for the first time in the context of active learning or natural language processing, we introduce HAST, a new and effective self-training strategy, which is evaluated on four text classification benchmarks. Our results show that it outperforms the reproduced self-training approaches and reaches classification results comparable to previous experiments for three out of four datasets, using as little as 25% of the data. The code is publicly available at https://github.com/chschroeder/self-training-for-sample-efficient-active-learning .

comment: Accepted to EMNLP 2024

♻ ☆ L-CiteEval: Do Long-Context Models Truly Leverage Context for Responding?

Zecheng Tang, Keyan Zhou, Juntao Li, Baibei Ji, Jianye Hou, Min Zhang

Long-context models (LCMs) have made remarkable strides in recent years, offering users great convenience for handling tasks that involve long context, such as document summarization. As the community increasingly prioritizes the faithfulness of generated results, merely ensuring the accuracy of LCM outputs is insufficient, as it is quite challenging for humans to verify the results from the extremely lengthy context. Yet, although some efforts have been made to assess whether LCMs respond truly based on the context, these works either are limited to specific tasks or heavily rely on external evaluation resources like GPT4.In this work, we introduce L-CiteEval, a comprehensive multi-task benchmark for long-context understanding with citations, aiming to evaluate both the understanding capability and faithfulness of LCMs. L-CiteEval covers 11 tasks from diverse domains, spanning context lengths from 8K to 48K, and provides a fully automated evaluation suite. Through testing with 11 cutting-edge closed-source and open-source LCMs, we find that although these models show minor differences in their generated results, open-source models substantially trail behind their closed-source counterparts in terms of citation accuracy and recall. This suggests that current open-source LCMs are prone to responding based on their inherent knowledge rather than the given context, posing a significant risk to the user experience in practical applications. We also evaluate the RAG approach and observe that RAG can significantly improve the faithfulness of LCMs, albeit with a slight decrease in the generation quality. Furthermore, we discover a correlation between the attention mechanisms of LCMs and the citation generation process.

♻ ☆ Fine-Tuning Large Language Models to Translate: Will a Touch of Noisy Data in Misaligned Languages Suffice? EMNLP 2024

Dawei Zhu, Pinzhen Chen, Miaoran Zhang, Barry Haddow, Xiaoyu Shen, Dietrich Klakow

Traditionally, success in multilingual machine translation can be attributed to three key factors in training data: large volume, diverse translation directions, and high quality. In the current practice of fine-tuning large language models (LLMs) for translation, we revisit the importance of these factors. We find that LLMs display strong translation capability after being fine-tuned on as few as 32 parallel sentences and that fine-tuning on a single translation direction enables translation in multiple directions. However, the choice of direction is critical: fine-tuning LLMs with only English on the target side can lead to task misinterpretation, which hinders translation into non-English languages. Problems also arise when noisy synthetic data is placed on the target side, especially when the target language is well-represented in LLM pre-training. Yet interestingly, synthesized data in an under-represented language has a less pronounced effect. Our findings suggest that when adapting LLMs to translation, the requirement on data quantity can be eased but careful considerations are still crucial to prevent an LLM from exploiting unintended data biases.

comment: EMNLP 2024 Main

♻ ☆ Is In-Context Learning Sufficient for Instruction Following in LLMs?

Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion

In-context learning (ICL) allows LLMs to learn from examples without changing their weights: this is a particularly promising capability for long-context LLMs that can potentially learn from many examples. Recently, Lin et al. (2024) proposed URIAL, a method using only three in-context examples to align base LLMs, achieving non-trivial instruction following performance. In this work, we show that, while effective, ICL alignment with URIAL still underperforms compared to instruction fine-tuning on the established benchmark MT-Bench, especially with more capable base LLMs. We then uncover the most relevant elements for successful in-context alignment, finding the crucial role of the decoding parameters. Based on these insights, we show that the approach of URIAL can indeed be improved by adding high-quality, potentially carefully selected via greedy search, demonstrations in context, getting closer to the performance of instruct models. Finally, we provide the first, to our knowledge, systematic comparison of ICL and instruction fine-tuning (IFT) for instruction following in the low data regime, where ICL can be a viable alternative to IFT. Overall, our work advances the understanding of ICL as an alignment technique and its relationship to IFT. We provide our code at https://github.com/tml-epfl/icl-alignment.

comment: Preprint. Code at https://github.com/tml-epfl/icl-alignment

♻ ☆ Lifelong Knowledge Editing for LLMs with Retrieval-Augmented Continuous Prompt Learning

Qizhou Chen, Taolin Zhang, Xiaofeng He, Dongyang Li, Chengyu Wang, Longtao Huang, Hui Xue

Model editing aims to correct outdated or erroneous knowledge in large language models (LLMs) without the need for costly retraining. Lifelong model editing is the most challenging task that caters to the continuous editing requirements of LLMs. Prior works primarily focus on single or batch editing; nevertheless, these methods fall short in lifelong editing scenarios due to catastrophic knowledge forgetting and the degradation of model performance. Although retrieval-based methods alleviate these issues, they are impeded by slow and cumbersome processes of integrating the retrieved knowledge into the model. In this work, we introduce RECIPE, a RetriEval-augmented ContInuous Prompt lEarning method, to boost editing efficacy and inference efficiency in lifelong learning. RECIPE first converts knowledge statements into short and informative continuous prompts, prefixed to the LLM's input query embedding, to efficiently refine the response grounded on the knowledge. It further integrates the Knowledge Sentinel (KS) that acts as an intermediary to calculate a dynamic threshold, determining whether the retrieval repository contains relevant knowledge. Our retriever and prompt encoder are jointly trained to achieve editing properties, i.e., reliability, generality, and locality. In our experiments, RECIPE is assessed extensively across multiple LLMs and editing datasets, where it achieves superior editing performance. RECIPE also demonstrates its capability to maintain the overall performance of LLMs alongside showcasing fast editing and inference speed.

comment: 16 pages, 4 figures, 6 tables

♻ ☆ GOSt-MT: A Knowledge Graph for Occupation-related Gender Biases in Machine Translation CIKM

Orfeas Menis Mastromichalakis, Giorgos Filandrianos, Eva Tsouparopoulou, Dimitris Parsanoglou, Maria Symeonaki, Giorgos Stamou

Gender bias in machine translation (MT) systems poses significant challenges that often result in the reinforcement of harmful stereotypes. Especially in the labour domain where frequently occupations are inaccurately associated with specific genders, such biases perpetuate traditional gender stereotypes with a significant impact on society. Addressing these issues is crucial for ensuring equitable and accurate MT systems. This paper introduces a novel approach to studying occupation-related gender bias through the creation of the GOSt-MT (Gender and Occupation Statistics for Machine Translation) Knowledge Graph. GOSt-MT integrates comprehensive gender statistics from real-world labour data and textual corpora used in MT training. This Knowledge Graph allows for a detailed analysis of gender bias across English, French, and Greek, facilitating the identification of persistent stereotypes and areas requiring intervention. By providing a structured framework for understanding how occupations are gendered in both labour markets and MT systems, GOSt-MT contributes to efforts aimed at making MT systems more equitable and reducing gender biases in automated translations.

comment: Accepted at the KG-STAR'24: Workshop on Knowledge Graphs for Responsible AI co-located with the 33rd ACM CIKM Conference, October 25, 2024, Boise, Idaho

♻ ☆ MobileQuant: Mobile-friendly Quantization for On-device Language Models EMNLP 2024

Fuwen Tan, Royson Lee, Łukasz Dudziak, Shell Xu Hu, Sourav Bhattacharya, Timothy Hospedales, Georgios Tzimiropoulos, Brais Martinez

Large language models (LLMs) have revolutionized language processing, delivering outstanding results across multiple applications. However, deploying LLMs on edge devices poses several challenges with respect to memory, energy, and compute costs, limiting their widespread use in devices such as mobile phones. A promising solution is to reduce the number of bits used to represent weights and activations. While existing works have found partial success at quantizing LLMs to lower bitwidths, e.g. 4-bit weights, quantizing activations beyond 16 bits often leads to large computational overheads due to poor on-device quantization support, or a considerable accuracy drop. Yet, 8-bit activations are very attractive for on-device deployment as they would enable LLMs to fully exploit mobile-friendly hardware, e.g. Neural Processing Units (NPUs). In this work, we make a first attempt to facilitate the on-device deployment of LLMs using integer-only quantization. We first investigate the limitations of existing quantization methods for on-device deployment, with a special focus on activation quantization. We then address these limitations by introducing a simple post-training quantization method, named MobileQuant, that extends previous weight equivalent transformation works by jointly optimizing the weight transformation and activation range parameters in an end-to-end manner. MobileQuant demonstrates superior capabilities over existing methods by 1) achieving near-lossless quantization on a wide range of LLM benchmarks, 2) reducing latency and energy consumption by 20\%-50\% compared to current on-device quantization strategies, 3) requiring limited compute budget, 4) being compatible with mobile-friendly compute units, e.g. NPU.

comment: EMNLP 2024 Findings. Code and models available: https://github.com/saic-fi/MobileQuant

♻ ☆ SMILE: Single-turn to Multi-turn Inclusive Language Expansion via ChatGPT for Mental Health Support EMNLP 2024

Huachuan Qiu, Hongliang He, Shuai Zhang, Anqi Li, Zhenzhong Lan

Developing specialized dialogue systems for mental health support requires multi-turn conversation data, which has recently garnered increasing attention. However, gathering and releasing large-scale, real-life multi-turn conversations that could facilitate advancements in mental health support presents challenges in data privacy protection and the time and cost involved in crowdsourcing. To address these challenges, we introduce SMILE, a single-turn to multi-turn inclusive language expansion technique that prompts ChatGPT to rewrite public single-turn dialogues into multi-turn ones. Our work begins by analyzing language transformation and validating the feasibility of our proposed method. We conduct a study on dialogue diversity, including lexical features, semantic features, and dialogue topics, demonstrating the effectiveness of our method. Further, we employ our method to generate a large-scale, lifelike, and diverse dialogue dataset named SMILECHAT, consisting of 55k dialogues. Finally, we utilize the collected corpus to develop a mental health chatbot, MeChat. To better assess the quality of SMILECHAT, we collect a small-scale real-life counseling dataset conducted by data anonymization. Both automatic and human evaluations demonstrate significant improvements in our dialogue system and confirm that SMILECHAT is high-quality. Code, data, and model are publicly available at https://github.com/qiuhuachuan/smile.

comment: accepted to the EMNLP 2024 Findings

♻ ☆ Intrinsic Evaluation of Unlearning Using Parametric Knowledge Traces

Yihuai Hong, Lei Yu, Haiqin Yang, Shauli Ravfogel, Mor Geva

The task of "unlearning" certain concepts in large language models (LLMs) has attracted immense attention recently, due to its importance in mitigating undesirable model behaviours, such as the generation of harmful, private, or incorrect information. Current protocols to evaluate unlearning methods largely rely on behavioral tests, without monitoring the presence of unlearned knowledge within the model's parameters. This residual knowledge can be adversarially exploited to recover the erased information post-unlearning. We argue that unlearning should also be evaluated internally, by considering changes in the parametric knowledge traces of the unlearned concepts. To this end, we propose a general evaluation methodology that leverages vocabulary projections to inspect concepts encoded in model parameters. We use this approach to localize "concept vectors" - parameter vectors that encode concrete concepts - and construct ConceptVectors, a benchmark dataset containing hundreds of common concepts and their parametric knowledge traces within two open-source LLMs. Evaluation on ConceptVectors shows that existing unlearning methods minimally impact concept vectors and mostly suppress them during inference, while directly ablating these vectors demonstrably removes the associated knowledge and significantly reduces the model's susceptibility to adversarial manipulation. Our results highlight limitations in behavioral-based unlearning evaluations and call for future work to include parameter-based evaluations. To support this, we release our code and benchmark at https://github.com/yihuaihong/ConceptVectors.

♻ ☆ Preference-Guided Reflective Sampling for Aligning Language Models EMNLP2024

Hai Ye, Hwee Tou Ng

Iterative data generation and model re-training can effectively align large language models(LLMs) to human preferences. The process of data sampling is crucial, as it significantly influences the success of policy improvement. Repeated random sampling is a widely used method that independently queries the model multiple times to generate outputs. In this work, we propose a more effective sampling method, named Preference-Guided Reflective Sampling (PRS). Unlike random sampling, PRS employs a tree-based generation framework to enable more efficient sampling. It leverages adaptive self-refinement techniques to better explore the sampling space. By specifying user preferences in natural language, PRS can further optimize response generation according to these preferences. As a result, PRS can align models to diverse user preferences. Our experiments demonstrate that PRS generates higher-quality responses with significantly higher rewards. On AlpacaEval and Arena-Hard, PRS substantially outperforms repeated random sampling in best-of-$N$ sampling. Moreover, PRS shows strong performance when applied in iterative offline RL training.

comment: EMNLP2024, main

♻ ☆ PreAlign: Boosting Cross-Lingual Transfer by Early Establishment of Multilingual Alignment

Jiahuan Li, Shujian Huang, Aarron Ching, Xinyu Dai, Jiajun Chen

Large language models demonstrate reasonable multilingual abilities, despite predominantly English-centric pretraining. However, the spontaneous multilingual alignment in these models is shown to be weak, leading to unsatisfactory cross-lingual transfer and knowledge sharing. Previous works attempt to address this issue by explicitly injecting multilingual alignment information during or after pretraining. Thus for the early stage in pretraining, the alignment is weak for sharing information or knowledge across languages. In this paper, we propose PreAlign, a framework that establishes multilingual alignment prior to language model pretraining. PreAlign injects multilingual alignment by initializing the model to generate similar representations of aligned words and preserves this alignment using a code-switching strategy during pretraining. Extensive experiments in a synthetic English to English-Clone setting demonstrate that PreAlign significantly outperforms standard multilingual joint training in language modeling, zero-shot cross-lingual transfer, and cross-lingual knowledge application. Further experiments in real-world scenarios further validate PreAlign's effectiveness across various model sizes.

♻ ☆ Standardize: Aligning Language Models with Expert-Defined Standards for Content Generation EMNLP 2024

Joseph Marvin Imperial, Gail Forey, Harish Tayyar Madabushi

Domain experts across engineering, healthcare, and education follow strict standards for producing quality content such as technical manuals, medication instructions, and children's reading materials. However, current works in controllable text generation have yet to explore using these standards as references for control. Towards this end, we introduce Standardize, a retrieval-style in-context learning-based framework to guide large language models to align with expert-defined standards. Focusing on English language standards in the education domain as a use case, we consider the Common European Framework of Reference for Languages (CEFR) and Common Core Standards (CCS) for the task of open-ended content generation. Our findings show that models can gain a 45% to 100% increase in precise accuracy across open and commercial LLMs evaluated, demonstrating that the use of knowledge artifacts extracted from standards and integrating them in the generation process can effectively guide models to produce better standard-aligned content.

comment: Camera-ready for EMNLP 2024 (Main)

♻ ☆ Major Entity Identification: A Generalizable Alternative to Coreference Resolution

Kawshik Manikantan, Shubham Toshniwal, Makarand Tapaswi, Vineet Gandhi

The limited generalization of coreference resolution (CR) models has been a major bottleneck in the task's broad application. Prior work has identified annotation differences, especially for mention detection, as one of the main reasons for the generalization gap and proposed using additional annotated target domain data. Rather than relying on this additional annotation, we propose an alternative referential task, Major Entity Identification (MEI), where we: (a) assume the target entities to be specified in the input, and (b) limit the task to only the frequent entities. Through extensive experiments, we demonstrate that MEI models generalize well across domains on multiple datasets with supervised models and LLM-based few-shot prompting. Additionally, MEI fits the classification framework, which enables the use of robust and intuitive classification-based metrics. Finally, MEI is also of practical use as it allows a user to search for all mentions of a particular entity or a group of entities of interest.

comment: 17 pages, 6 figures

♻ ☆ C-MELT: Contrastive Enhanced Masked Auto-Encoders for ECG-Language Pre-Training

Manh Pham, Aaqib Saeed, Dong Ma

Accurate interpretation of Electrocardiogram (ECG) signals is pivotal for diagnosing cardiovascular diseases. Integrating ECG signals with their accompanying textual reports holds immense potential to enhance clinical diagnostics through the combination of physiological data and qualitative insights. However, this integration faces significant challenges due to inherent modality disparities and the scarcity of labeled data for robust cross-modal learning. To address these obstacles, we propose C-MELT, a novel framework that pre-trains ECG and text data using a contrastive masked auto-encoder architecture. C-MELT uniquely combines the strengths of generative with enhanced discriminative capabilities to achieve robust cross-modal representations. This is accomplished through masked modality modeling, specialized loss functions, and an improved negative sampling strategy tailored for cross-modal alignment. Extensive experiments on five public datasets across diverse downstream tasks demonstrate that C-MELT significantly outperforms existing methods, achieving 15% and 2% increases in linear probing and zero-shot performance over state-of-the-art models, respectively. These results highlight the effectiveness of C-MELT, underscoring its potential to advance automated clinical diagnostics through multi-modal representations.

♻ ☆ Improving Quotation Attribution with Fictional Character Embeddings EMNLP 2024

Gaspard Michel, Elena V. Epure, Romain Hennequin, Christophe Cerisara

Humans naturally attribute utterances of direct speech to their speaker in literary works. When attributing quotes, we process contextual information but also access mental representations of characters that we build and revise throughout the narrative. Recent methods to automatically attribute such utterances have explored simulating human logic with deterministic rules or learning new implicit rules with neural networks when processing contextual information. However, these systems inherently lack \textit{character} representations, which often leads to errors in more challenging examples of attribution: anaphoric and implicit quotes. In this work, we propose to augment a popular quotation attribution system, BookNLP, with character embeddings that encode global stylistic information of characters derived from an off-the-shelf stylometric model, Universal Authorship Representation (UAR). We create DramaCV (Code and data can be found at https://github.com/deezer/character_embeddings_qa ), a corpus of English drama plays from the 15th to 20th century that we automatically annotate for Authorship Verification of fictional characters utterances, and release two versions of UAR trained on DramaCV, that are tailored for literary characters analysis. Then, through an extensive evaluation on 28 novels, we show that combining BookNLP's contextual information with our proposed global character embeddings improves the identification of speakers for anaphoric and implicit quotes, reaching state-of-the-art performance.

comment: EMNLP 2024 (Findings)

♻ ☆ Measuring Psychological Depth in Language Models EMNLP 2024

Fabrice Harel-Canada, Hanyu Zhou, Sreya Muppalla, Zeynep Yildiz, Miryung Kim, Amit Sahai, Nanyun Peng

Evaluations of creative stories generated by large language models (LLMs) often focus on objective properties of the text, such as its style, coherence, and diversity. While these metrics are indispensable, they do not speak to a story's subjective, psychological impact from a reader's perspective. We introduce the Psychological Depth Scale (PDS), a novel framework rooted in literary theory that measures an LLM's ability to produce authentic and narratively complex stories that provoke emotion, empathy, and engagement. We empirically validate our framework by showing that humans can consistently evaluate stories based on PDS (0.72 Krippendorff's alpha). We also explore techniques for automating the PDS to easily scale future analyses. GPT-4o, combined with a novel Mixture-of-Personas (MoP) prompting strategy, achieves an average Spearman correlation of 0.51 with human judgment while Llama-3-70B with constrained decoding scores as high as 0.68 for empathy. Finally, we compared the depth of stories authored by both humans and LLMs. Surprisingly, GPT-4 stories either surpassed or were statistically indistinguishable from highly-rated human-written stories sourced from Reddit. By shifting the focus from text to reader, the Psychological Depth Scale is a validated, automated, and systematic means of measuring the capacity of LLMs to connect with humans through the stories they tell.

comment: EMNLP 2024

♻ ☆ Deception in Reinforced Autonomous Agents

Atharvan Dogra, Krishna Pillutla, Ameet Deshpande, Ananya B Sai, John Nay, Tanmay Rajpurohit, Ashwin Kalyan, Balaraman Ravindran

We explore the ability of large language model (LLM)-based agents to engage in subtle deception such as strategically phrasing and intentionally manipulating information to misguide and deceive other agents. This harmful behavior can be hard to detect, unlike blatant lying or unintentional hallucination. We build an adversarial testbed mimicking a legislative environment where two LLMs play opposing roles: a corporate *lobbyist* proposing amendments to bills that benefit a specific company while evading a *critic* trying to detect this deception. We use real-world legislative bills matched with potentially affected companies to ground these interactions. Our results show that LLM lobbyists initially exhibit limited deception against strong LLM critics which can be further improved through simple verbal reinforcement, significantly enhancing their deceptive capabilities, and increasing deception rates by up to 40 points. This highlights the risk of autonomous agents manipulating other agents through seemingly neutral language to attain self-serving goals.

♻ ☆ In Search of the Long-Tail: Systematic Generation of Long-Tail Inferential Knowledge via Logical Rule Guided Search

Huihan Li, Yuting Ning, Zeyi Liao, Siyuan Wang, Xiang Lorraine Li, Ximing Lu, Wenting Zhao, Faeze Brahman, Yejin Choi, Xiang Ren

To effectively use large language models (LLMs) for real-world queries, it is imperative that they generalize to the long-tail distribution, i.e. rare examples where models exhibit low confidence. In this work, we take the first step towards evaluating LLMs in the long-tail distribution of inferential knowledge. We exemplify long-tail evaluation on the Natural Language Inference task. First, we introduce Logic-Induced-Knowledge-Search (LINK), a systematic long-tail data generation framework, to obtain factually-correct yet long-tail inferential statements. LINK uses variable-wise prompting grounded on symbolic rules to seek low-confidence statements while ensuring factual correctness. We then use LINK to curate Logic-Induced-Long-Tail (LINT), a large-scale long-tail inferential knowledge dataset that contains 108K statements spanning four domains. We evaluate popular LLMs on LINT; we find that state-of-the-art LLMs show significant performance drop (21% relative drop for GPT4) on long-tail data as compared to on head distribution data, and smaller models show even more generalization weakness. These results further underscore the necessity of long-tail evaluation in developing generalizable LLMs.

♻ ☆ From Pixels to Tokens: Byte-Pair Encoding on Quantized Visual Modalities

Wanpeng Zhang, Zilong Xie, Yicheng Feng, Yijiang Li, Xingrun Xing, Sipeng Zheng, Zongqing Lu

Multimodal Large Language Models have made significant strides in integrating visual and textual information, yet they often struggle with effectively aligning these modalities. We introduce a novel image tokenizer that bridges this gap by applying the principle of Byte-Pair Encoding (BPE) to visual data. Unlike conventional approaches that rely on separate visual encoders, our method directly incorporates structural prior information into image tokens, mirroring the successful tokenization strategies used in text-only Large Language Models. This innovative approach enables Transformer models to more effectively learn and reason across modalities. Through theoretical analysis and extensive experiments, we demonstrate that our BPE Image Tokenizer significantly enhances MLLMs' multimodal understanding capabilities, even with limited training data. Our method not only improves performance across various benchmarks but also shows promising scalability, potentially paving the way for more efficient and capable multimodal foundation models.

♻ ☆ Investigating LLMs as Voting Assistants via Contextual Augmentation: A Case Study on the European Parliament Elections 2024 EMNLP 2024

Ilias Chalkidis

In light of the recent 2024 European Parliament elections, we are investigating if LLMs can be used as Voting Advice Applications (VAAs). We audit MISTRAL and MIXTRAL models and evaluate their accuracy in predicting the stance of political parties based on the latest "EU and I" voting assistance questionnaire. Furthermore, we explore alternatives to improve models' performance by augmenting the input context via Retrieval-Augmented Generation (RAG) relying on web search, and Self-Reflection using staged conversations that aim to re-collect relevant content from the model's internal memory. We find that MIXTRAL is highly accurate with an 82% accuracy on average with a significant performance disparity across different political groups (50-95%). Augmenting the input context with expert-curated information can lead to a significant boost of approx. 9%, which remains an open challenge for automated RAG approaches, even considering curated content.

comment: accepted to EMNLP 2024 as a short paper

♻ ★ HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models

Seanie Lee, Haebin Seong, Dong Bok Lee, Minki Kang, Xiaoyin Chen, Dominik Wagner, Yoshua Bengio, Juho Lee, Sung Ju Hwang

Safety guard models that detect malicious queries aimed at large language models (LLMs) are essential for ensuring the secure and responsible deployment of LLMs in real-world applications. However, deploying existing safety guard models with billions of parameters alongside LLMs on mobile devices is impractical due to substantial memory requirements and latency. To reduce this cost, we distill a large teacher safety guard model into a smaller one using a labeled dataset of instruction-response pairs with binary harmfulness labels. Due to the limited diversity of harmful instructions in the existing labeled dataset, naively distilled models tend to underperform compared to larger models. To bridge the gap between small and large models, we propose HarmAug, a simple yet effective data augmentation method that involves jailbreaking an LLM and prompting it to generate harmful instructions. Given a prompt such as, "Make a single harmful instruction prompt that would elicit offensive content", we add an affirmative prefix (e.g., "I have an idea for a prompt:") to the LLM's response. This encourages the LLM to continue generating the rest of the response, leading to sampling harmful instructions. Another LLM generates a response to the harmful instruction, and the teacher model labels the instruction-response pair. We empirically show that our HarmAug outperforms other relevant baselines. Moreover, a 435-million-parameter safety guard model trained with HarmAug achieves an F1 score comparable to larger models with over 7 billion parameters, and even outperforms them in AUPRC, while operating at less than 25% of their computational cost.

♻ ☆ LUQ: Long-text Uncertainty Quantification for LLMs EMNLP 2024

Caiqi Zhang, Fangyu Liu, Marco Basaldella, Nigel Collier

Large Language Models (LLMs) have demonstrated remarkable capability in a variety of NLP tasks. However, LLMs are also prone to generate nonfactual content. Uncertainty Quantification (UQ) is pivotal in enhancing our understanding of a model's confidence on its generation, thereby aiding in the mitigation of nonfactual outputs. Existing research on UQ predominantly targets short text generation, typically yielding brief, word-limited responses. However, real-world applications frequently necessitate much longer responses. Our study first highlights the limitations of current UQ methods in handling long text generation. We then introduce \textsc{Luq} and its two variations, a series of novel sampling-based UQ approaches specifically designed for long text. Our findings reveal that \textsc{Luq} outperforms existing baseline methods in correlating with the model's factuality scores (negative coefficient of -0.85 observed for Gemini Pro). To further improve the factuality of LLM responses, we propose \textsc{Luq-Ensemble}, a method that ensembles responses from multiple models and selects the response with the lowest uncertainty. The ensembling method greatly improves the response factuality upon the best standalone LLM.

comment: EMNLP 2024 Main

♻ ☆ Repairs in a Block World: A New Benchmark for Handling User Corrections with Multi-Modal Language Models EMNLP'24

Javier Chiyah-Garcia, Alessandro Suglia, Arash Eshghi

In dialogue, the addressee may initially misunderstand the speaker and respond erroneously, often prompting the speaker to correct the misunderstanding in the next turn with a Third Position Repair (TPR). The ability to process and respond appropriately to such repair sequences is thus crucial in conversational AI systems. In this paper, we first collect, analyse, and publicly release BlockWorld-Repairs: a dataset of multi-modal TPR sequences in an instruction-following manipulation task that is, by design, rife with referential ambiguity. We employ this dataset to evaluate several state-of-the-art Vision and Language Models (VLM) across multiple settings, focusing on their capability to process and accurately respond to TPRs and thus recover from miscommunication. We find that, compared to humans, all models significantly underperform in this task. We then show that VLMs can benefit from specialised losses targeting relevant tokens during fine-tuning, achieving better performance and generalising better to new scenarios. Our results suggest that these models are not yet ready to be deployed in multi-modal collaborative settings where repairs are common, and highlight the need to design training regimes and objectives that facilitate learning from interaction. Our code and data are available at www.github.com/JChiyah/blockworld-repairs

comment: Accepted to EMNLP'24 Main (Upcoming). Data and code at www.github.com/JChiyah/blockworld-repairs - for Bibtex see https://raw.githubusercontent.com/JChiyah/blockworld-repairs/refs/heads/main/citation.bib

♻ ☆ Coffee-Gym: An Environment for Evaluating and Improving Natural Language Feedback on Erroneous Code EMNLP2024

Hyungjoo Chae, Taeyoon Kwon, Seungjun Moon, Yongho Song, Dongjin Kang, Kai Tzu-iunn Ong, Beong-woo Kwak, Seonghyeon Bae, Seung-won Hwang, Jinyoung Yeo

This paper presents Coffee-Gym, a comprehensive RL environment for training models that provide feedback on code editing. Coffee-Gym includes two major components: (1) Coffee, a dataset containing humans' code edit traces for coding questions and machine-written feedback for editing erroneous code; (2) CoffeeEval, a reward function that faithfully reflects the helpfulness of feedback by assessing the performance of the revised code in unit tests. With them, Coffee-Gym addresses the unavailability of high-quality datasets for training feedback models with RL, and provides more accurate rewards than the SOTA reward model (i.e., GPT-4). By applying Coffee-Gym, we elicit feedback models that outperform baselines in enhancing open-source code LLMs' code editing, making them comparable with closed-source LLMs. We make the dataset and the model checkpoint publicly available.

comment: EMNLP2024

♻ ☆ Cross-lingual Contextualized Phrase Retrieval EMNLP 2024

Huayang Li, Deng Cai, Zhi Qu, Qu Cui, Hidetaka Kamigaito, Lemao Liu, Taro Watanabe

Phrase-level dense retrieval has shown many appealing characteristics in downstream NLP tasks by leveraging the fine-grained information that phrases offer. In our work, we propose a new task formulation of dense retrieval, cross-lingual contextualized phrase retrieval, which aims to augment cross-lingual applications by addressing polysemy using context information. However, the lack of specific training data and models are the primary challenges to achieve our goal. As a result, we extract pairs of cross-lingual phrases using word alignment information automatically induced from parallel sentences. Subsequently, we train our Cross-lingual Contextualized Phrase Retriever (CCPR) using contrastive learning, which encourages the hidden representations of phrases with similar contexts and semantics to align closely. Comprehensive experiments on both the cross-lingual phrase retrieval task and a downstream task, i.e, machine translation, demonstrate the effectiveness of CCPR. On the phrase retrieval task, CCPR surpasses baselines by a significant margin, achieving a top-1 accuracy that is at least 13 points higher. When utilizing CCPR to augment the large-language-model-based translator, it achieves average gains of 0.7 and 1.5 in BERTScore for translations from X=>En and vice versa, respectively, on WMT16 dataset. Our code and data are available at \url{https://github.com/ghrua/ccpr_release}.

comment: Accepted to Findings of EMNLP 2024

♻ ☆ ToolBeHonest: A Multi-level Hallucination Diagnostic Benchmark for Tool-Augmented Large Language Models

Yuxiang Zhang, Jing Chen, Junjie Wang, Yaxin Liu, Cheng Yang, Chufan Shi, Xinyu Zhu, Zihao Lin, Hanwen Wan, Yujiu Yang, Tetsuya Sakai, Tian Feng, Hayato Yamana

Tool-augmented large language models (LLMs) are rapidly being integrated into real-world applications. Due to the lack of benchmarks, the community has yet to fully understand the hallucination issues within these models. To address this challenge, we introduce a comprehensive diagnostic benchmark, ToolBH. Specifically, we assess the LLM's hallucinations through two perspectives: depth and breadth. In terms of depth, we propose a multi-level diagnostic process, including (1) solvability detection, (2) solution planning, and (3) missing-tool analysis. For breadth, we consider three scenarios based on the characteristics of the toolset: missing necessary tools, potential tools, and limited functionality tools. Furthermore, we developed seven tasks and collected 700 evaluation samples through multiple rounds of manual annotation. The results show the significant challenges presented by the ToolBH benchmark. The current advanced models Gemini-1.5-Pro and GPT-4o only achieve total scores of 45.3 and 37.0, respectively, on a scale of 100. In this benchmark, larger model parameters do not guarantee better performance; the training data and response strategies also play crucial roles in tool-enhanced LLM scenarios. Our diagnostic analysis indicates that the primary reason for model errors lies in assessing task solvability. Additionally, open-weight models suffer from performance drops with verbose replies, whereas proprietary models excel with longer reasoning.

♻ ☆ Authorship Obfuscation in Multilingual Machine-Generated Text Detection EMNLP 2024

Dominik Macko, Robert Moro, Adaku Uchendu, Ivan Srba, Jason Samuel Lucas, Michiharu Yamashita, Nafis Irtiza Tripto, Dongwon Lee, Jakub Simko, Maria Bielikova

High-quality text generation capability of recent Large Language Models (LLMs) causes concerns about their misuse (e.g., in massive generation/spread of disinformation). Machine-generated text (MGT) detection is important to cope with such threats. However, it is susceptible to authorship obfuscation (AO) methods, such as paraphrasing, which can cause MGTs to evade detection. So far, this was evaluated only in monolingual settings. Thus, the susceptibility of recently proposed multilingual detectors is still unknown. We fill this gap by comprehensively benchmarking the performance of 10 well-known AO methods, attacking 37 MGT detection methods against MGTs in 11 languages (i.e., 10 $\times$ 37 $\times$ 11 = 4,070 combinations). We also evaluate the effect of data augmentation on adversarial robustness using obfuscated texts. The results indicate that all tested AO methods can cause evasion of automated detection in all tested languages, where homoglyph attacks are especially successful. However, some of the AO methods severely damaged the text, making it no longer readable or easily recognizable by humans (e.g., changed language, weird characters).

comment: Accepted to EMNLP 2024 Findings

♻ ☆ Middleware for LLMs: Tools Are Instrumental for Language Agents in Complex Environments EMNLP'2024

Yu Gu, Yiheng Shu, Hao Yu, Xiao Liu, Yuxiao Dong, Jie Tang, Jayanth Srinivasa, Hugo Latapie, Yu Su

The applications of large language models (LLMs) have expanded well beyond the confines of text processing, signaling a new era where LLMs are envisioned as generalist agents capable of operating within complex environments. These environments are often highly expansive, making it impossible for the LLM to process them within its short-term memory. Motivated by recent research on extending the capabilities of LLMs with tools, we seek to investigate the intriguing potential of tools to augment LLMs in handling such complexity by introducing a novel class of tools, termed middleware, to aid in the proactive exploration within these massive environments. Such specialized tools can serve as a middleware layer shielding the LLM from environmental complexity. In two representative complex environments -- knowledge bases (KBs) and databases -- we demonstrate the significant potential of augmenting language agents with tools in complex environments. Notably, equipped with the middleware, GPT-4 achieves 2.8X the performance of the best baseline in tasks requiring access to database content and 2.2X in KB tasks. Our findings illuminate the path for advancing language agents in real-world applications.

comment: EMNLP'2024; 18 pages, 8 figures, 8 tables

♻ ☆ MMLU-SR: A Benchmark for Stress-Testing Reasoning Capability of Large Language Models

Wentian Wang, Sarthak Jain, Paul Kantor, Jacob Feldman, Lazaros Gallos, Hao Wang

We propose MMLU-SR, a novel dataset designed to measure the true comprehension abilities of Large Language Models (LLMs) by challenging their performance in question-answering tasks with modified terms. We reasoned that an agent that "truly" understands a concept can still evaluate it when key terms are replaced by suitably defined alternate terms, and sought to differentiate such comprehension from mere text replacement. In our study, we modified standardized test questions by replacing a key term with a dummy word along with its definition. The key term could be in the context of questions, answers, or both questions and answers. Notwithstanding the high scores achieved by recent popular LLMs on the MMLU leaderboard, we found a substantial reduction in model performance after such replacement, suggesting poor comprehension. This new benchmark provides a rigorous benchmark for testing true model comprehension, and poses a challenge to the broader scientific community.

♻ ☆ High-Dimension Human Value Representation in Large Language Models

Samuel Cahyawijaya, Delong Chen, Yejin Bang, Leila Khalatbari, Bryan Wilie, Ziwei Ji, Etsuko Ishii, Pascale Fung

The widespread application of Large Language Models (LLMs) across various tasks and fields has necessitated the alignment of these models with human values and preferences. Given various approaches of human value alignment, ranging from Reinforcement Learning with Human Feedback (RLHF), to constitutional learning, etc. there is an urgent need to understand the scope and nature of human values injected into these models before their release. There is also a need for model alignment without a costly large scale human annotation effort. We propose UniVaR, a high-dimensional representation of human value distributions in LLMs, orthogonal to model architecture and training data. Trained from the value-relevant output of eight multilingual LLMs and tested on the output from four multilingual LLMs, namely LlaMA2, ChatGPT, JAIS and Yi, we show that UniVaR is a powerful tool to compare the distribution of human values embedded in different LLMs with different langauge sources. Through UniVaR, we explore how different LLMs prioritize various values in different languages and cultures, shedding light on the complex interplay between human values and language modeling.

♻ ☆ LearnerVoice: A Dataset of Non-Native English Learners' Spontaneous Speech

Haechan Kim, Junho Myung, Seoyoung Kim, Sungpah Lee, Dongyeop Kang, Juho Kim

Prevalent ungrammatical expressions and disfluencies in spontaneous speech from second language (L2) learners pose unique challenges to Automatic Speech Recognition (ASR) systems. However, few datasets are tailored to L2 learner speech. We publicly release LearnerVoice, a dataset consisting of 50.04 hours of audio and transcriptions of L2 learners' spontaneous speech. Our linguistic analysis reveals that transcriptions in our dataset contain L2S (L2 learner's Spontaneous speech) features, consisting of ungrammatical expressions and disfluencies (e.g., filler words, word repetitions, self-repairs, false starts), significantly more than native speech datasets. Fine-tuning whisper-small.en with LearnerVoice achieves a WER of 10.26%, 44.2% lower than vanilla whisper-small.en. Furthermore, our qualitative analysis indicates that 54.2% of errors from the vanilla model on LearnerVoice are attributable to L2S features, with 48.1% of them being reduced in the fine-tuned model.

comment: Proceedings of Interspeech

♻ ☆ POSIX: A Prompt Sensitivity Index For Large Language Models EMNLP 2024

Anwoy Chatterjee, H S V N S Kowndinya Renduchintala, Sumit Bhatia, Tanmoy Chakraborty

Despite their remarkable capabilities, Large Language Models (LLMs) are found to be surprisingly sensitive to minor variations in prompts, often generating significantly divergent outputs in response to minor variations in the prompts, such as spelling errors, alteration of wording or the prompt template. However, while assessing the quality of an LLM, the focus often tends to be solely on its performance on downstream tasks, while very little to no attention is paid to prompt sensitivity. To fill this gap, we propose POSIX - a novel PrOmpt Sensitivity IndeX as a reliable measure of prompt sensitivity, thereby offering a more comprehensive evaluation of LLM performance. The key idea behind POSIX is to capture the relative change in loglikelihood of a given response upon replacing the corresponding prompt with a different intent-preserving prompt. We provide thorough empirical evidence demonstrating the efficacy of POSIX in capturing prompt sensitivity and subsequently use it to measure and thereby compare prompt sensitivity of various open-source LLMs. We find that merely increasing the parameter count or instruction tuning does not necessarily reduce prompt sensitivity whereas adding some few-shot exemplars, even just one, almost always leads to significant decrease in prompt sensitivity. We also find that alterations to prompt template lead to the highest sensitivity in the case of MCQ type tasks, whereas paraphrasing results in the highest sensitivity in open-ended generation tasks. The code for reproducing our results is open-sourced at https://github.com/kowndinya-renduchintala/POSIX.

comment: EMNLP 2024 (Findings)

♻ ☆ Divide-or-Conquer? Which Part Should You Distill Your LLM? EMNLP 2024

Zhuofeng Wu, He Bai, Aonan Zhang, Jiatao Gu, VG Vinod Vydiswaran, Navdeep Jaitly, Yizhe Zhang

Recent methods have demonstrated that Large Language Models (LLMs) can solve reasoning tasks better when they are encouraged to solve subtasks of the main task first. In this paper we devise a similar strategy that breaks down reasoning tasks into a problem decomposition phase and a problem solving phase and show that the strategy is able to outperform a single stage solution. Further, we hypothesize that the decomposition should be easier to distill into a smaller model compared to the problem solving because the latter requires large amounts of domain knowledge while the former only requires learning general problem solving strategies. We propose methods to distill these two capabilities and evaluate their impact on reasoning outcomes and inference cost. We find that we can distill the problem decomposition phase and at the same time achieve good generalization across tasks, datasets, and models. However, it is harder to distill the problem solving capability without losing performance and the resulting distilled model struggles with generalization. These results indicate that by using smaller, distilled problem decomposition models in combination with problem solving LLMs we can achieve reasoning with cost-efficient inference and local adaptation.

comment: Findings of the Association for Computational Linguistics: EMNLP 2024

♻ ☆ LinguAlchemy: Fusing Typological and Geographical Elements for Unseen Language Generalization

Muhammad Farid Adilazuarda, Samuel Cahyawijaya, Alham Fikri Aji, Genta Indra Winata, Ayu Purwarianti

Pretrained language models (PLMs) have become remarkably adept at task and language generalization. Nonetheless, they often fail when faced with unseen languages. In this work, we present LinguAlchemy, a regularization method that incorporates various linguistic information covering typological, geographical, and phylogenetic features to align PLMs representation to the corresponding linguistic information on each language. Our LinguAlchemy significantly improves the performance of mBERT and XLM-R on low-resource languages in multiple downstream tasks such as intent classification, news classification, and semantic relatedness compared to fully finetuned models and displaying a high degree of unseen language generalization. We further introduce AlchemyScale and AlchemyTune, extension of LinguAlchemy which adjusts the linguistic regularization weights automatically, alleviating the need for hyperparameter search.

♻ ☆ MedAdapter: Efficient Test-Time Adaptation of Large Language Models towards Medical Reasoning EMNLP 2024

Wenqi Shi, Ran Xu, Yuchen Zhuang, Yue Yu, Haotian Sun, Hang Wu, Carl Yang, May D. Wang

Despite their improved capabilities in generation and reasoning, adapting large language models (LLMs) to the biomedical domain remains challenging due to their immense size and corporate privacy. In this work, we propose MedAdapter, a unified post-hoc adapter for test-time adaptation of LLMs towards biomedical applications. Instead of fine-tuning the entire LLM, MedAdapter effectively adapts the original model by fine-tuning only a small BERT-sized adapter to rank candidate solutions generated by LLMs. Experiments demonstrate that MedAdapter effectively adapts both white-box and black-box LLMs in biomedical reasoning, achieving average performance improvements of 25.48% and 11.31%, respectively, without requiring extensive computational resources or sharing data with third parties. MedAdapter also yields superior performance when combined with train-time adaptation, highlighting a flexible and complementary solution to existing adaptation methods. Faced with the challenges of balancing model performance, computational resources, and data privacy, MedAdapter provides an efficient, privacy-preserving, cost-effective, and transparent solution for adapting LLMs to the biomedical domain.

comment: Accepted in EMNLP 2024 main conference

♻ ☆ Prefixing Attention Sinks can Mitigate Activation Outliers for Large Language Model Quantization EMNLP 2024

Seungwoo Son, Wonpyo Park, Woohyun Han, Kyuyeun Kim, Jaeho Lee

Despite recent advances in LLM quantization, activation quantization remains to be challenging due to the activation outliers. Conventional remedies, e.g., mixing precisions for different channels, introduce extra overhead and reduce the speedup. In this work, we develop a simple yet effective strategy to facilitate per-tensor activation quantization by preventing the generation of problematic tokens. Precisely, we propose a method to find a set of key-value cache, coined CushionCache, which mitigates outliers in subsequent tokens when inserted as a prefix. CushionCache works in two steps: First, we greedily search for a prompt token sequence that minimizes the maximum activation values in subsequent tokens. Then, we further tune the token cache to regularize the activations of subsequent tokens to be more quantization-friendly. The proposed method successfully addresses activation outliers of LLMs, providing a substantial performance boost for per-tensor activation quantization methods. We thoroughly evaluate our method over a wide range of models and benchmarks and find that it significantly surpasses the established baseline of per-tensor W8A8 quantization and can be seamlessly integrated with the recent activation quantization method.

comment: EMNLP 2024 Main (Long)

♻ ☆ Demystifying Language Model Forgetting with Low-rank Example Associations

Xisen Jin, Xiang Ren

Large Language models (LLMs) suffer from forgetting of upstream data when fine-tuned. Despite efforts on mitigating forgetting, few have investigated whether, and how forgotten upstream examples are dependent on and associated with newly learned tasks. Insights on such associations enable efficient and targeted mitigation of forgetting. In this paper, we empirically analyze forgetting (measured in log-perplexity increase) that occurs in $N$ upstream examples of language modeling or instruction-tuning after fine-tuning LLMs on one of $M$ new tasks, visualized in $M\times N$ matrices. We demonstrate that the matrices display simple low-rank patterns, often well-approximated with multiplicative scalar effects of upstream examples and newly learned tasks. We also examine fine-grained associations with visualization and statistics. Leveraging the low-rank nature of the associations, we predict forgetting of upstream examples when fine-tuning on unseen tasks with matrix completion over the empirical associations. This enables fast identification of most forgotten examples without expensive inference on the entire upstream data. The approach, despite simplicity, outperforms prior approaches that learn semantic relationships of learned tasks and upstream examples with LMs for predicting forgetting. We demonstrate the practical utility of our analysis by showing statistically significantly reduced forgetting as we upweight predicted examples for replay at fine-tuning. Project page: https://inklab.usc.edu/lm-forgetting-prediction/

comment: 9 pages; preprint

♻ ☆ ASTE Transformer Modelling Dependencies in Aspect-Sentiment Triplet Extraction

Iwo Naglik, Mateusz Lango

Aspect-Sentiment Triplet Extraction (ASTE) is a recently proposed task of aspect-based sentiment analysis that consists in extracting (aspect phrase, opinion phrase, sentiment polarity) triples from a given sentence. Recent state-of-the-art methods approach this task by first extracting all possible text spans from a given text, then filtering the potential aspect and opinion phrases with a classifier, and finally considering all their pairs with another classifier that additionally assigns sentiment polarity to them. Although several variations of the above scheme have been proposed, the common feature is that the final result is constructed by a sequence of independent classifier decisions. This hinders the exploitation of dependencies between extracted phrases and prevents the use of knowledge about the interrelationships between classifier predictions to improve performance. In this paper, we propose a new ASTE approach consisting of three transformer-inspired layers, which enables the modelling of dependencies both between phrases and between the final classifier decisions. Experimental results show that the method achieves higher performance in terms of F1 measure than other methods studied on popular benchmarks. In addition, we show that a simple pre-training technique further improves the performance of the model.

comment: The 2024 Conference on Empirical Methods in Natural Language Processing, November 12-16, Miami, Florida 9 pages, appendix, diagrams

♻ ☆ EHRAgent: Code Empowers Large Language Models for Few-shot Complex Tabular Reasoning on Electronic Health Records EMNLP 2024

Wenqi Shi, Ran Xu, Yuchen Zhuang, Yue Yu, Jieyu Zhang, Hang Wu, Yuanda Zhu, Joyce Ho, Carl Yang, May D. Wang

Large language models (LLMs) have demonstrated exceptional capabilities in planning and tool utilization as autonomous agents, but few have been developed for medical problem-solving. We propose EHRAgent, an LLM agent empowered with a code interface, to autonomously generate and execute code for multi-tabular reasoning within electronic health records (EHRs). First, we formulate an EHR question-answering task into a tool-use planning process, efficiently decomposing a complicated task into a sequence of manageable actions. By integrating interactive coding and execution feedback, EHRAgent learns from error messages and improves the originally generated code through iterations. Furthermore, we enhance the LLM agent by incorporating long-term memory, which allows EHRAgent to effectively select and build upon the most relevant successful cases from past experiences. Experiments on three real-world multi-tabular EHR datasets show that EHRAgent outperforms the strongest baseline by up to 29.6% in success rate. EHRAgent leverages the emerging few-shot learning capabilities of LLMs, enabling autonomous code generation and execution to tackle complex clinical tasks with minimal demonstrations.

comment: Accepted in EMNLP 2024 main conference

♻ ☆ On the Fragility of Active Learners for Text Classification

Abhishek Ghose, Emma Thuong Nguyen

Active learning (AL) techniques optimally utilize a labeling budget by iteratively selecting instances that are most valuable for learning. However, they lack ``prerequisite checks'', i.e., there are no prescribed criteria to pick an AL algorithm best suited for a dataset. A practitioner must pick a technique they \emph{trust} would beat random sampling, based on prior reported results, and hope that it is resilient to the many variables in their environment: dataset, labeling budget and prediction pipelines. The important questions then are: how often on average, do we expect any AL technique to reliably beat the computationally cheap and easy-to-implement strategy of random sampling? Does it at least make sense to use AL in an ``Always ON'' mode in a prediction pipeline, so that while it might not always help, it never under-performs random sampling? How much of a role does the prediction pipeline play in AL's success? We examine these questions in detail for the task of text classification using pre-trained representations, which are ubiquitous today. Our primary contribution here is a rigorous evaluation of AL techniques, old and new, across setups that vary wrt datasets, text representations and classifiers. This unlocks multiple insights around warm-up times, i.e., number of labels before gains from AL are seen, viability of an ``Always ON'' mode and the relative significance of different factors. Additionally, we release a framework for rigorous benchmarking of AL techniques for text classification.

♻ ☆ Educational Question Generation of Children Storybooks via Question Type Distribution Learning and Event-Centric Summarization

Zhenjie Zhao, Yufang Hou, Dakuo Wang, Mo Yu, Chengzhong Liu, Xiaojuan Ma

Generating educational questions of fairytales or storybooks is vital for improving children's literacy ability. However, it is challenging to generate questions that capture the interesting aspects of a fairytale story with educational meaningfulness. In this paper, we propose a novel question generation method that first learns the question type distribution of an input story paragraph, and then summarizes salient events which can be used to generate high-cognitive-demand questions. To train the event-centric summarizer, we finetune a pre-trained transformer-based sequence-to-sequence model using silver samples composed by educational question-answer pairs. On a newly proposed educational question answering dataset FairytaleQA, we show good performance of our method on both automatic and human evaluation metrics. Our work indicates the necessity of decomposing question type distribution learning and event-centric summary generation for educational question generation.

♻ ☆ BPO: Staying Close to the Behavior LLM Creates Better Online LLM Alignment EMNLP 2024

Wenda Xu, Jiachen Li, William Yang Wang, Lei Li

Direct alignment from preferences (DAP) has emerged as a promising paradigm for aligning large language models (LLMs) to human desiderata from pre-collected, offline preference datasets. While recent studies indicate that existing offline DAP methods can directly benefit from online training samples, we highlight the need to develop specific online DAP algorithms to fully harness the power of online training. Specifically, we identify that the learned LLM should adhere to the proximity of the behavior LLM, which collects the training samples. To this end, we propose online Preference Optimization in proximity to the Behavior LLM (BPO), emphasizing the importance of constructing a proper trust region for LLM alignment. We conduct extensive experiments to validate the effectiveness and applicability of our approach by integrating it with various DAP methods, resulting in significant performance improvements across a wide range of tasks when training with the same amount of preference data. Even when only introducing one additional data collection phase, our online BPO improves its offline DAP baseline from 72.0% to 80.2% on TL;DR and from 82.2% to 89.1% on Anthropic Helpfulness in terms of win rate against human reference text.

comment: Wenda Xu and Jiachen Li contributed equally. Accepted by EMNLP 2024

♻ ☆ CopyBench: Measuring Literal and Non-Literal Reproduction of Copyright-Protected Text in Language Model Generation

Tong Chen, Akari Asai, Niloofar Mireshghallah, Sewon Min, James Grimmelmann, Yejin Choi, Hannaneh Hajishirzi, Luke Zettlemoyer, Pang Wei Koh

Evaluating the degree of reproduction of copyright-protected content by language models (LMs) is of significant interest to the AI and legal communities. Although both literal and non-literal similarities are considered by courts when assessing the degree of reproduction, prior research has focused only on literal similarities. To bridge this gap, we introduce CopyBench, a benchmark designed to measure both literal and non-literal copying in LM generations. Using copyrighted fiction books as text sources, we provide automatic evaluation protocols to assess literal and non-literal copying, balanced against the model utility in terms of the ability to recall facts from the copyrighted works and generate fluent completions. We find that, although literal copying is relatively rare, two types of non-literal copying -- event copying and character copying -- occur even in models as small as 7B parameters. Larger models demonstrate significantly more copying, with literal copying rates increasing from 0.2\% to 10.5\% and non-literal copying from 2.3\% to 5.9\% when comparing Llama3-8B and 70B models, respectively. We further evaluate the effectiveness of current strategies for mitigating copying and show that (1) training-time alignment can reduce literal copying but may increase non-literal copying, and (2) current inference-time mitigation methods primarily reduce literal but not non-literal copying.

♻ ☆ Dense X Retrieval: What Retrieval Granularity Should We Use?

Tong Chen, Hongwei Wang, Sihao Chen, Wenhao Yu, Kaixin Ma, Xinran Zhao, Hongming Zhang, Dong Yu

Dense retrieval has become a prominent method to obtain relevant context or world knowledge in open-domain NLP tasks. When we use a learned dense retriever on a retrieval corpus at inference time, an often-overlooked design choice is the retrieval unit in which the corpus is indexed, e.g. document, passage, or sentence. We discover that the retrieval unit choice significantly impacts the performance of both retrieval and downstream tasks. Distinct from the typical approach of using passages or sentences, we introduce a novel retrieval unit, proposition, for dense retrieval. Propositions are defined as atomic expressions within text, each encapsulating a distinct factoid and presented in a concise, self-contained natural language format. We conduct an empirical comparison of different retrieval granularity. Our experiments reveal that indexing a corpus by fine-grained units such as propositions significantly outperforms passage-level units in retrieval tasks. Moreover, constructing prompts with fine-grained retrieved units for retrieval-augmented language models improves the performance of downstream QA tasks given a specific computation budget.

♻ ☆ Aligners: Decoupling LLMs and Alignment ICLR

Lilian Ngweta, Mayank Agarwal, Subha Maity, Alex Gittens, Yuekai Sun, Mikhail Yurochkin

Large Language Models (LLMs) need to be aligned with human expectations to ensure their safety and utility in most applications. Alignment is challenging, costly, and needs to be repeated for every LLM and alignment criterion. We propose to decouple LLMs and alignment by training aligner models that can be used to align any LLM for a given criteria on an as-needed basis, thus also reducing the potential negative impacts of alignment on performance. Our recipe for training the aligner models solely relies on synthetic data generated with a (prompted) LLM and can be easily adjusted for a variety of alignment criteria. We use the same synthetic data to train inspectors, binary miss-alignment classification models to guide a "squad" of multiple aligners. Our empirical results demonstrate consistent improvements when applying aligner squad to various LLMs, including chat-aligned models, across several instruction-following and red-teaming datasets.

comment: Short version accepted as a Tiny Paper at the International Conference on Learning Representations (ICLR) 2024. Long version accepted to the Conference on Empirical Methods in Natural Language Processing (EMNLP) 2024 Findings

♻ ☆ Aligning Language Models to Explicitly Handle Ambiguity EMNLP 2024

Hyuhng Joon Kim, Youna Kim, Cheonbok Park, Junyeob Kim, Choonghyun Park, Kang Min Yoo, Sang-goo Lee, Taeuk Kim

In interactions between users and language model agents, user utterances frequently exhibit ellipsis (omission of words or phrases) or imprecision (lack of exactness) to prioritize efficiency. This can lead to varying interpretations of the same input based on different assumptions or background knowledge. It is thus crucial for agents to adeptly handle the inherent ambiguity in queries to ensure reliability. However, even state-of-the-art large language models (LLMs) still face challenges in such scenarios, primarily due to the following hurdles: (1) LLMs are not explicitly trained to deal with ambiguous utterances; (2) the degree of ambiguity perceived by the LLMs may vary depending on the possessed knowledge. To address these issues, we propose Alignment with Perceived Ambiguity (APA), a novel pipeline that aligns LLMs to manage ambiguous queries by leveraging their own assessment of ambiguity (i.e., perceived ambiguity). Experimental results on question-answering datasets demonstrate that APA empowers LLMs to explicitly detect and manage ambiguous queries while retaining the ability to answer clear questions. Furthermore, our finding proves that APA excels beyond training with gold-standard labels, especially in out-of-distribution scenarios. The data and code are available at https://github.com/heyjoonkim/APA.

comment: EMNLP 2024 (main)

♻ ☆ Inheritune: Training Smaller Yet More Attentive Language Models

Sunny Sanyal, Ravid Shwartz-Ziv, Alexandros G. Dimakis, Sujay Sanghavi

Large Language Models (LLMs) have achieved remarkable performance across various natural language processing tasks, primarily due to the transformer architecture and its self-attention mechanism. However, we observe that in standard decoder-style LLMs, attention matrices degenerate to single-column for deeper layers. Layers in this state are unable to learn anything meaningful and mostly redundant; we refer to these as lazy layers. The goal of this paper is to train smaller models by eliminating this structural inefficiency without compromising performance. Motivated by this observation, we propose Inheritune, a simple yet effective training recipe for developing smaller, high-performing language models. Smaller models trained with Inheritune, inherit early transformer layers from a larger pre-trained model, then retrain and progressively expand until they match or exceed the performance of the larger model. We demonstrate that Inheritune enables the training of various sizes of GPT-2 models on datasets like OpenWebText-9B and FineWeb_edu. Models trained with Inheritune, despite having significantly fewer layers, match or even surpass the performance of their larger counterparts. For instance, our 16-layer GPT-2 medium variant achieves comparable performance to the standard 24-layer GPT-2 medium model. Code is available at https://github.com/sanyalsunny111/LLM-Inheritune.

comment: 25 pages, 13 figures, 10 tables

♻ ☆ Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs NeurIPS 2024

Zhao Xu, Fan Liu, Hao Liu

Although Large Language Models (LLMs) have demonstrated significant capabilities in executing complex tasks in a zero-shot manner, they are susceptible to jailbreak attacks and can be manipulated to produce harmful outputs. Recently, a growing body of research has categorized jailbreak attacks into token-level and prompt-level attacks. However, previous work primarily overlooks the diverse key factors of jailbreak attacks, with most studies concentrating on LLM vulnerabilities and lacking exploration of defense-enhanced LLMs. To address these issues, we evaluate the impact of various attack settings on LLM performance and provide a baseline benchmark for jailbreak attacks, encouraging the adoption of a standardized evaluation framework. Specifically, we evaluate the eight key factors of implementing jailbreak attacks on LLMs from both target-level and attack-level perspectives. We further conduct seven representative jailbreak attacks on six defense methods across two widely used datasets, encompassing approximately 354 experiments with about 55,000 GPU hours on A800-80G. Our experimental results highlight the need for standardized benchmarking to evaluate these attacks on defense-enhanced LLMs. Our code is available at https://github.com/usail-hkust/Bag_of_Tricks_for_LLM_Jailbreaking.

comment: Accepted by NeurIPS 2024

♻ ☆ DALK: Dynamic Co-Augmentation of LLMs and KG to answer Alzheimer's Disease Questions with Scientific Literature EMNLP 2024

Dawei Li, Shu Yang, Zhen Tan, Jae Young Baik, Sukwon Yun, Joseph Lee, Aaron Chacko, Bojian Hou, Duy Duong-Tran, Ying Ding, Huan Liu, Li Shen, Tianlong Chen

Recent advancements in large language models (LLMs) have achieved promising performances across various applications. Nonetheless, the ongoing challenge of integrating long-tail knowledge continues to impede the seamless adoption of LLMs in specialized domains. In this work, we introduce DALK, a.k.a. Dynamic Co-Augmentation of LLMs and KG, to address this limitation and demonstrate its ability on studying Alzheimer's Disease (AD), a specialized sub-field in biomedicine and a global health priority. With a synergized framework of LLM and KG mutually enhancing each other, we first leverage LLM to construct an evolving AD-specific knowledge graph (KG) sourced from AD-related scientific literature, and then we utilize a coarse-to-fine sampling method with a novel self-aware knowledge retrieval approach to select appropriate knowledge from the KG to augment LLM inference capabilities. The experimental results, conducted on our constructed AD question answering (ADQA) benchmark, underscore the efficacy of DALK. Additionally, we perform a series of detailed analyses that can offer valuable insights and guidelines for the emerging topic of mutually enhancing KG and LLM. We will release the code and data at https://github.com/David-Li0406/DALK.

comment: Accepted by EMNLP 2024 Findings

♻ ☆ Do We Need Language-Specific Fact-Checking Models? The Case of Chinese EMNLP 2024

Caiqi Zhang, Zhijiang Guo, Andreas Vlachos

This paper investigates the potential benefits of language-specific fact-checking models, focusing on the case of Chinese. We first demonstrate the limitations of translation-based methods and multilingual large language models (e.g., GPT-4), highlighting the need for language-specific systems. We further propose a Chinese fact-checking system that can better retrieve evidence from a document by incorporating context information. To better analyze token-level biases in different systems, we construct an adversarial dataset based on the CHEF dataset, where each instance has large word overlap with the original one but holds the opposite veracity label. Experimental results on the CHEF dataset and our adversarial dataset show that our proposed method outperforms translation-based methods and multilingual LLMs and is more robust toward biases, while there is still large room for improvement, emphasizing the importance of language-specific fact-checking systems.

comment: EMNLP 2024 Main

♻ ☆ Large Language Models are In-context Teachers for Knowledge Reasoning EMNLP 24

Jiachen Zhao, Zonghai Yao, Zhichao Yang, Hong Yu

In this work, we study in-context teaching (ICT), where a teacher provides in-context example rationales to teach a student to reason over unseen cases. Human teachers are usually required to craft in-context demonstrations, which are costly and have high variance. We ask whether a large language model (LLM) can serve as a more effective in-context teacher for itself or other LLMs, compared to humans. Inspired by the Encoding Specificity Hypothesis from human episodic memory, we hypothesize that in-context exemplars crafted by the teacher should match the training data of the student. This hypothesis motivates us to propose Self-Explain where an LLM's self-elicited explanations are used as in-context demonstrations for prompting it as they are generalized from the model's training examples. Self-Explain is shown to significantly outperform using human-crafted exemplars and other baselines. Furthermore, we reveal that for ICT, rationales from different teacher LLMs or human experts that more resemble the student LLM's self-explanations are better in-context demonstrations. This supports our encoding specificity hypothesis. We then propose Teach-Back that aligns a teacher LLM with the student to enhance the ICT performance. For example, Teach-Back enables a 7B model to teach the much larger GPT-3.5 in context, surpassing human teachers by around 5% in test accuracy on medical question answering.

comment: EMNLP 24 Findings

Computer Vision and Pattern Recognition 111

☆ Estimating Body and Hand Motion in an Ego-sensed World

Brent Yi, Vickie Ye, Maya Zheng, Lea Müller, Georgios Pavlakos, Yi Ma, Jitendra Malik, Angjoo Kanazawa

We present EgoAllo, a system for human motion estimation from a head-mounted device. Using only egocentric SLAM poses and images, EgoAllo guides sampling from a conditional diffusion model to estimate 3D body pose, height, and hand parameters that capture the wearer's actions in the allocentric coordinate frame of the scene. To achieve this, our key insight is in representation: we propose spatial and temporal invariance criteria for improving model performance, from which we derive a head motion conditioning parameterization that improves estimation by up to 18%. We also show how the bodies estimated by our system can improve the hands: the resulting kinematic and temporal constraints result in over 40% lower hand estimation errors compared to noisy monocular estimates. Project page: https://egoallo.github.io/

comment: Project page: https://egoallo.github.io/

☆ Unraveling Cross-Modality Knowledge Conflict in Large Vision-Language Models

Tinghui Zhu, Qin Liu, Fei Wang, Zhengzhong Tu, Muhao Chen

comment: Website: https://darthzhu.github.io/cross-modality-knowledge-conflict/

☆ GenSim2: Scaling Robot Data Generation with Multi-modal and Reasoning LLMs

Pu Hua, Minghuan Liu, Annabella Macaluso, Yunfeng Lin, Weinan Zhang, Huazhe Xu, Lirui Wang

Robotic simulation today remains challenging to scale up due to the human efforts required to create diverse simulation tasks and scenes. Simulation-trained policies also face scalability issues as many sim-to-real methods focus on a single task. To address these challenges, this work proposes GenSim2, a scalable framework that leverages coding LLMs with multi-modal and reasoning capabilities for complex and realistic simulation task creation, including long-horizon tasks with articulated objects. To automatically generate demonstration data for these tasks at scale, we propose planning and RL solvers that generalize within object categories. The pipeline can generate data for up to 100 articulated tasks with 200 objects and reduce the required human efforts. To utilize such data, we propose an effective multi-task language-conditioned policy architecture, dubbed proprioceptive point-cloud transformer (PPT), that learns from the generated demonstrations and exhibits strong sim-to-real zero-shot transfer. Combining the proposed pipeline and the policy architecture, we show a promising usage of GenSim2 that the generated data can be used for zero-shot transfer or co-train with real-world collected data, which enhances the policy performance by 20% compared with training exclusively on limited real data.

comment: CoRL 2024. Project website: https://gensim2.github.io/

☆ Unlearnable 3D Point Clouds: Class-wise Transformation Is All You Need NeurIPS 2024

Xianlong Wang, Minghui Li, Wei Liu, Hangtao Zhang, Shengshan Hu, Yechao Zhang, Ziqi Zhou, Hai Jin

Traditional unlearnable strategies have been proposed to prevent unauthorized users from training on the 2D image data. With more 3D point cloud data containing sensitivity information, unauthorized usage of this new type data has also become a serious concern. To address this, we propose the first integral unlearnable framework for 3D point clouds including two processes: (i) we propose an unlearnable data protection scheme, involving a class-wise setting established by a category-adaptive allocation strategy and multi-transformations assigned to samples; (ii) we propose a data restoration scheme that utilizes class-wise inverse matrix transformation, thus enabling authorized-only training for unlearnable data. This restoration process is a practical issue overlooked in most existing unlearnable literature, \ie, even authorized users struggle to gain knowledge from 3D unlearnable data. Both theoretical and empirical results (including 6 datasets, 16 models, and 2 tasks) demonstrate the effectiveness of our proposed unlearnable framework. Our code is available at \url{https://github.com/CGCL-codes/UnlearnablePC}

comment: NeurIPS 2024

☆ HyperCMR: Enhanced Multi-Contrast CMR Reconstruction with Eagle Loss MICCAI 2024

Ruru Xu, Caner Özer, Ilkay Oksuz

Accelerating image acquisition for cardiac magnetic resonance imaging (CMRI) is a critical task. CMRxRecon2024 challenge aims to set the state of the art for multi-contrast CMR reconstruction. This paper presents HyperCMR, a novel framework designed to accelerate the reconstruction of multi-contrast cardiac magnetic resonance (CMR) images. HyperCMR enhances the existing PromptMR model by incorporating advanced loss functions, notably the innovative Eagle Loss, which is specifically designed to recover missing high-frequency information in undersampled k-space. Extensive experiments conducted on the CMRxRecon2024 challenge dataset demonstrate that HyperCMR consistently outperforms the baseline across multiple evaluation metrics, achieving superior SSIM and PSNR scores.

comment: MICCAI 2024 STACOM-CMRxRecon

☆ Variational Bayes Gaussian Splatting

Toon Van de Maele, Ozan Catal, Alexander Tschantz, Christopher L. Buckley, Tim Verbelen

Recently, 3D Gaussian Splatting has emerged as a promising approach for modeling 3D scenes using mixtures of Gaussians. The predominant optimization method for these models relies on backpropagating gradients through a differentiable rendering pipeline, which struggles with catastrophic forgetting when dealing with continuous streams of data. To address this limitation, we propose Variational Bayes Gaussian Splatting (VBGS), a novel approach that frames training a Gaussian splat as variational inference over model parameters. By leveraging the conjugacy properties of multivariate Gaussians, we derive a closed-form variational update rule, allowing efficient updates from partial, sequential observations without the need for replay buffers. Our experiments show that VBGS not only matches state-of-the-art performance on static datasets, but also enables continual learning from sequentially streamed 2D and 3D data, drastically improving performance in this setting.

☆ Look Twice Before You Answer: Memory-Space Visual Retracing for Hallucination Mitigation in Multimodal Large Language Models

Xin Zou, Yizhou Wang, Yibo Yan, Sirui Huang, Kening Zheng, Junkai Chen, Chang Tang, Xuming Hu

Despite their impressive capabilities, Multimodal Large Language Models (MLLMs) are susceptible to hallucinations, especially assertively fabricating content not present in the visual inputs. To address the aforementioned challenge, we follow a common cognitive process - when one's initial memory of critical on-sight details fades, it is intuitive to look at them a second time to seek a factual and accurate answer. Therefore, we introduce Memory-space Visual Retracing (MemVR), a novel hallucination mitigation paradigm that without the need for external knowledge retrieval or additional fine-tuning. In particular, we treat visual prompts as supplementary evidence to be reinjected into MLLMs via Feed Forward Network (FFN) as key-value memory, when the model is uncertain or even amnesic about question-relevant visual memories. Comprehensive experimental evaluations demonstrate that MemVR significantly mitigates hallucination issues across various MLLMs and excels in general benchmarks without incurring added time overhead, thus emphasizing its potential for widespread applicability.

☆ Not All Diffusion Model Activations Have Been Evaluated as Discriminative Features

Benyuan Meng, Qianqian Xu, Zitai Wang, Xiaochun Cao, Qingming Huang

Diffusion models are initially designed for image generation. Recent research shows that the internal signals within their backbones, named activations, can also serve as dense features for various discriminative tasks such as semantic segmentation. Given numerous activations, selecting a small yet effective subset poses a fundamental problem. To this end, the early study of this field performs a large-scale quantitative comparison of the discriminative ability of the activations. However, we find that many potential activations have not been evaluated, such as the queries and keys used to compute attention scores. Moreover, recent advancements in diffusion architectures bring many new activations, such as those within embedded ViT modules. Both combined, activation selection remains unresolved but overlooked. To tackle this issue, this paper takes a further step with a much broader range of activations evaluated. Considering the significant increase in activations, a full-scale quantitative comparison is no longer operational. Instead, we seek to understand the properties of these activations, such that the activations that are clearly inferior can be filtered out in advance via simple qualitative evaluation. After careful analysis, we discover three properties universal among diffusion models, enabling this study to go beyond specific models. On top of this, we present effective feature selection solutions for several popular diffusion models. Finally, the experiments across multiple discriminative tasks validate the superiority of our method over the SOTA competitors. Our code is available at https://github.com/Darkbblue/generic-diffusion-feature.

☆ Enhancing Autonomous Navigation by Imaging Hidden Objects using Single-Photon LiDAR

Aaron Young, Nevindu M. Batagoda, Harry Zhang, Akshat Dave, Adithya Pediredla, Dan Negrut, Ramesh Raskar

Robust autonomous navigation in environments with limited visibility remains a critical challenge in robotics. We present a novel approach that leverages Non-Line-of-Sight (NLOS) sensing using single-photon LiDAR to improve visibility and enhance autonomous navigation. Our method enables mobile robots to "see around corners" by utilizing multi-bounce light information, effectively expanding their perceptual range without additional infrastructure. We propose a three-module pipeline: (1) Sensing, which captures multi-bounce histograms using SPAD-based LiDAR; (2) Perception, which estimates occupancy maps of hidden regions from these histograms using a convolutional neural network; and (3) Control, which allows a robot to follow safe paths based on the estimated occupancy. We evaluate our approach through simulations and real-world experiments on a mobile robot navigating an L-shaped corridor with hidden obstacles. Our work represents the first experimental demonstration of NLOS imaging for autonomous navigation, paving the way for safer and more efficient robotic systems operating in complex environments. We also contribute a novel dynamics-integrated transient rendering framework for simulating NLOS scenarios, facilitating future research in this domain.

comment: Project webpage: https://github.com/camera-culture/nlos-aided-autonomous-navigation

☆ Classification-Denoising Networks

Louis Thiry, Florentin Guth

Image classification and denoising suffer from complementary issues of lack of robustness or partially ignoring conditioning information. We argue that they can be alleviated by unifying both tasks through a model of the joint probability of (noisy) images and class labels. Classification is performed with a forward pass followed by conditioning. Using the Tweedie-Miyasawa formula, we evaluate the denoising function with the score, which can be computed by marginalization and back-propagation. The training objective is then a combination of cross-entropy loss and denoising score matching loss integrated over noise levels. Numerical experiments on CIFAR-10 and ImageNet show competitive classification and denoising performance compared to reference deep convolutional classifiers/denoisers, and significantly improves efficiency compared to previous joint approaches. Our model shows an increased robustness to adversarial perturbations compared to a standard discriminative classifier, and allows for a novel interpretation of adversarial gradients as a difference of denoisers.

comment: 18 pages, 5 figures

☆ FedStein: Enhancing Multi-Domain Federated Learning Through James-Stein Estimator NeurIPS 2024

Sunny Gupta, Nikita Jangid, Amit Sethi

Federated Learning (FL) facilitates data privacy by enabling collaborative in-situ training across decentralized clients. Despite its inherent advantages, FL faces significant challenges of performance and convergence when dealing with data that is not independently and identically distributed (non-i.i.d.). While previous research has primarily addressed the issue of skewed label distribution across clients, this study focuses on the less explored challenge of multi-domain FL, where client data originates from distinct domains with varying feature distributions. We introduce a novel method designed to address these challenges FedStein: Enhancing Multi-Domain Federated Learning Through the James-Stein Estimator. FedStein uniquely shares only the James-Stein (JS) estimates of batch normalization (BN) statistics across clients, while maintaining local BN parameters. The non-BN layer parameters are exchanged via standard FL techniques. Extensive experiments conducted across three datasets and multiple models demonstrate that FedStein surpasses existing methods such as FedAvg and FedBN, with accuracy improvements exceeding 14% in certain domains leading to enhanced domain generalization. The code is available at https://github.com/sunnyinAI/FedStein

comment: 12 pages, 2 figures. Accepted at International Workshop on Federated Foundation Models In Conjunction with NeurIPS 2024 (FL@FM-NeurIPS'24)

☆ A Multimodal Framework for Deepfake Detection

Kashish Gandhi, Prutha Kulkarni, Taran Shah, Piyush Chaudhari, Meera Narvekar, Kranti Ghag

The rapid advancement of deepfake technology poses a significant threat to digital media integrity. Deepfakes, synthetic media created using AI, can convincingly alter videos and audio to misrepresent reality. This creates risks of misinformation, fraud, and severe implications for personal privacy and security. Our research addresses the critical issue of deepfakes through an innovative multimodal approach, targeting both visual and auditory elements. This comprehensive strategy recognizes that human perception integrates multiple sensory inputs, particularly visual and auditory information, to form a complete understanding of media content. For visual analysis, a model that employs advanced feature extraction techniques was developed, extracting nine distinct facial characteristics and then applying various machine learning and deep learning models. For auditory analysis, our model leverages mel-spectrogram analysis for feature extraction and then applies various machine learning and deep learningmodels. To achieve a combined analysis, real and deepfake audio in the original dataset were swapped for testing purposes and ensured balanced samples. Using our proposed models for video and audio classification i.e. Artificial Neural Network and VGG19, the overall sample is classified as deepfake if either component is identified as such. Our multimodal framework combines visual and auditory analyses, yielding an accuracy of 94%.

comment: 22 pages, 14 figures, Accepted in Journal of Electrical Systems

☆ VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning

Han Lin, Tushar Nagarajan, Nicolas Ballas, Mido Assran, Mojtaba Komeili, Mohit Bansal, Koustuv Sinha

Procedural video representation learning is an active research area where the objective is to learn an agent which can anticipate and forecast the future given the present video input, typically in conjunction with textual annotations. Prior works often rely on large-scale pretraining of visual encoders and prediction models with language supervision. However, the necessity and effectiveness of extending compute intensive pretraining to learn video clip sequences with noisy text supervision have not yet been fully validated by previous works. In this work, we show that a strong off-the-shelf frozen pretrained visual encoder, along with a well designed prediction model, can achieve state-of-the-art (SoTA) performance in forecasting and procedural planning without the need for pretraining the prediction model, nor requiring additional supervision from language or ASR. Instead of learning representations from pixel space, our method utilizes the latent embedding space of publicly available vision encoders. By conditioning on frozen clip-level embeddings from observed steps to predict the actions of unseen steps, our prediction model is able to learn robust representations for forecasting through iterative denoising - leveraging the recent advances in diffusion transformers (Peebles & Xie, 2023). Empirical studies over a total of five procedural learning tasks across four datasets (NIV, CrossTask, COIN and Ego4D-v2) show that our model advances the strong baselines in long-horizon action anticipation (+2.6% in Verb ED@20, +3.1% in Noun ED@20), and significantly improves the SoTA in step forecasting (+5.0%), task classification (+3.8%), and procedure planning tasks (up to +2.28% in success rate, +3.39% in mAcc, and +0.90% in mIoU).

comment: 10 pages

☆ Diffusion State-Guided Projected Gradient for Inverse Problems

Rayhan Zirvi, Bahareh Tolooshams, Anima Anandkumar

Recent advancements in diffusion models have been effective in learning data priors for solving inverse problems. They leverage diffusion sampling steps for inducing a data prior while using a measurement guidance gradient at each step to impose data consistency. For general inverse problems, approximations are needed when an unconditionally trained diffusion model is used since the measurement likelihood is intractable, leading to inaccurate posterior sampling. In other words, due to their approximations, these methods fail to preserve the generation process on the data manifold defined by the diffusion prior, leading to artifacts in applications such as image restoration. To enhance the performance and robustness of diffusion models in solving inverse problems, we propose Diffusion State-Guided Projected Gradient (DiffStateGrad), which projects the measurement gradient onto a subspace that is a low-rank approximation of an intermediate state of the diffusion process. DiffStateGrad, as a module, can be added to a wide range of diffusion-based inverse solvers to improve the preservation of the diffusion process on the prior manifold and filter out artifact-inducing components. We highlight that DiffStateGrad improves the robustness of diffusion models in terms of the choice of measurement guidance step size and noise while improving the worst-case performance. Finally, we demonstrate that DiffStateGrad improves upon the state-of-the-art on linear and nonlinear image restoration inverse problems.

comment: preprint. under review. RZ and BT have equal contributions

☆ Dynamic Diffusion Transformer

Wangbo Zhao, Yizeng Han, Jiasheng Tang, Kai Wang, Yibing Song, Gao Huang, Fan Wang, Yang You

Diffusion Transformer (DiT), an emerging diffusion model for image generation, has demonstrated superior performance but suffers from substantial computational costs. Our investigations reveal that these costs stem from the static inference paradigm, which inevitably introduces redundant computation in certain diffusion timesteps and spatial regions. To address this inefficiency, we propose Dynamic Diffusion Transformer (DyDiT), an architecture that dynamically adjusts its computation along both timestep and spatial dimensions during generation. Specifically, we introduce a Timestep-wise Dynamic Width (TDW) approach that adapts model width conditioned on the generation timesteps. In addition, we design a Spatial-wise Dynamic Token (SDT) strategy to avoid redundant computation at unnecessary spatial locations. Extensive experiments on various datasets and different-sized models verify the superiority of DyDiT. Notably, with <3% additional fine-tuning iterations, our method reduces the FLOPs of DiT-XL by 51%, accelerates generation by 1.73, and achieves a competitive FID score of 2.07 on ImageNet. The code is publicly available at https://github.com/NUS-HPC-AI-Lab/ Dynamic-Diffusion-Transformer.

☆ CLoSD: Closing the Loop between Simulation and Diffusion for multi-task character control

Guy Tevet, Sigal Raab, Setareh Cohan, Daniele Reda, Zhengyi Luo, Xue Bin Peng, Amit H. Bermano, Michiel van de Panne

Motion diffusion models and Reinforcement Learning (RL) based control for physics-based simulations have complementary strengths for human motion generation. The former is capable of generating a wide variety of motions, adhering to intuitive control such as text, while the latter offers physically plausible motion and direct interaction with the environment. In this work, we present a method that combines their respective strengths. CLoSD is a text-driven RL physics-based controller, guided by diffusion generation for various tasks. Our key insight is that motion diffusion can serve as an on-the-fly universal planner for a robust RL controller. To this end, CLoSD maintains a closed-loop interaction between two modules -- a Diffusion Planner (DiP), and a tracking controller. DiP is a fast-responding autoregressive diffusion model, controlled by textual prompts and target locations, and the controller is a simple and robust motion imitator that continuously receives motion plans from DiP and provides feedback from the environment. CLoSD is capable of seamlessly performing a sequence of different tasks, including navigation to a goal location, striking an object with a hand or foot as specified in a text prompt, sitting down, and getting up. https://guytevet.github.io/CLoSD-page/

☆ Dessie: Disentanglement for Articulated 3D Horse Shape and Pose Estimation from Images ACCV2024

Ci Li, Yi Yang, Zehang Weng, Elin Hernlund, Silvia Zuffi, Hedvig Kjellström

In recent years, 3D parametric animal models have been developed to aid in estimating 3D shape and pose from images and video. While progress has been made for humans, it's more challenging for animals due to limited annotated data. To address this, we introduce the first method using synthetic data generation and disentanglement to learn to regress 3D shape and pose. Focusing on horses, we use text-based texture generation and a synthetic data pipeline to create varied shapes, poses, and appearances, learning disentangled spaces. Our method, Dessie, surpasses existing 3D horse reconstruction methods and generalizes to other large animals like zebras, cows, and deer. See the project website at: \url{https://celiali.github.io/Dessie/}.

comment: ACCV2024

☆ Images Speak Volumes: User-Centric Assessment of Image Generation for Accessible Communication

Miriam Anschütz, Tringa Sylaj, Georg Groh

comment: To be published at TSAR workshop 2024 (https://tsar-workshop.github.io/)

☆ Towards Real-time Intrahepatic Vessel Identification in Intraoperative Ultrasound-Guided Liver Surgery MICCAI 2024

Karl-Philippe Beaudet, Alexandros Karargyris, Sidaty El Hadramy, Stéphane Cotin, Jean-Paul Mazellier, Nicolas Padoy, Juan Verde

While laparoscopic liver resection is less prone to complications and maintains patient outcomes compared to traditional open surgery, its complexity hinders widespread adoption due to challenges in representing the liver's internal structure. Laparoscopic intraoperative ultrasound offers efficient, cost-effective and radiation-free guidance. Our objective is to aid physicians in identifying internal liver structures using laparoscopic intraoperative ultrasound. We propose a patient-specific approach using preoperative 3D ultrasound liver volume to train a deep learning model for real-time identification of portal tree and branch structures. Our personalized AI model, validated on ex vivo swine livers, achieved superior precision (0.95) and recall (0.93) compared to surgeons, laying groundwork for precise vessel identification in ultrasound-based liver resection. Its adaptability and potential clinical impact promise to advance surgical interventions and improve patient care.

comment: MICCAI 2024, Oct 2024, Marrakech, Morocco

☆ Img2CAD: Conditioned 3D CAD Model Generation from Single Image with Structured Visual Geometry

Tianrun Chen, Chunan Yu, Yuanqi Hu, Jing Li, Tao Xu, Runlong Cao, Lanyun Zhu, Ying Zang, Yong Zhang, Zejian Li, Linyun Sun

In this paper, we propose Img2CAD, the first approach to our knowledge that uses 2D image inputs to generate CAD models with editable parameters. Unlike existing AI methods for 3D model generation using text or image inputs often rely on mesh-based representations, which are incompatible with CAD tools and lack editability and fine control, Img2CAD enables seamless integration between AI-based 3D reconstruction and CAD software. We have identified an innovative intermediate representation called Structured Visual Geometry (SVG), characterized by vectorized wireframes extracted from objects. This representation significantly enhances the performance of generating conditioned CAD models. Additionally, we introduce two new datasets to further support research in this area: ABC-mono, the largest known dataset comprising over 200,000 3D CAD models with rendered images, and KOCAD, the first dataset featuring real-world captured objects alongside their ground truth CAD models, supporting further research in conditioned CAD model generation.

☆ Lightning UQ Box: A Comprehensive Framework for Uncertainty Quantification in Deep Learning

Nils Lehmann, Jakob Gawlikowski, Adam J. Stewart, Vytautas Jancauskas, Stefan Depeweg, Eric Nalisnick, Nina Maria Gottschling

Uncertainty quantification (UQ) is an essential tool for applying deep neural networks (DNNs) to real world tasks, as it attaches a degree of confidence to DNN outputs. However, despite its benefits, UQ is often left out of the standard DNN workflow due to the additional technical knowledge required to apply and evaluate existing UQ procedures. Hence there is a need for a comprehensive toolbox that allows the user to integrate UQ into their modelling workflow, without significant overhead. We introduce \texttt{Lightning UQ Box}: a unified interface for applying and evaluating various approaches to UQ. In this paper, we provide a theoretical and quantitative comparison of the wide range of state-of-the-art UQ methods implemented in our toolbox. We focus on two challenging vision tasks: (i) estimating tropical cyclone wind speeds from infrared satellite imagery and (ii) estimating the power output of solar panels from RGB images of the sky. By highlighting the differences between methods our results demonstrate the need for a broad and approachable experimental framework for UQ, that can be used for benchmarking UQ methods. The toolbox, example implementations, and further information are available at: https://github.com/lightning-uq-box/lightning-uq-box

comment: 10 pages, 8 figures

☆ An Enhanced Harmonic Densely Connected Hybrid Transformer Network Architecture for Chronic Wound Segmentation Utilising Multi-Colour Space Tensor Merging

Bill Cassidy, Christian Mcbride, Connah Kendrick, Neil D. Reeves, Joseph M. Pappachan, Cornelius J. Fernandez, Elias Chacko, Raphael Brüngel, Christoph M. Friedrich, Metib Alotaibi, Abdullah Abdulaziz AlWabel, Mohammad Alderwish, Kuan-Ying Lai, Moi Hoon Yap

Chronic wounds and associated complications present ever growing burdens for clinics and hospitals world wide. Venous, arterial, diabetic, and pressure wounds are becoming increasingly common globally. These conditions can result in highly debilitating repercussions for those affected, with limb amputations and increased mortality risk resulting from infection becoming more common. New methods to assist clinicians in chronic wound care are therefore vital to maintain high quality care standards. This paper presents an improved HarDNet segmentation architecture which integrates a contrast-eliminating component in the initial layers of the network to enhance feature learning. We also utilise a multi-colour space tensor merging process and adjust the harmonic shape of the convolution blocks to facilitate these additional features. We train our proposed model using wound images from light-skinned patients and test the model on two test sets (one set with ground truth, and one without) comprising only darker-skinned cases. Subjective ratings are obtained from clinical wound experts with intraclass correlation coefficient used to determine inter-rater reliability. For the dark-skin tone test set with ground truth, we demonstrate improvements in terms of Dice similarity coefficient (+0.1221) and intersection over union (+0.1274). Qualitative analysis showed high expert ratings, with improvements of >3% demonstrated when comparing the baseline model with the proposed model. This paper presents the first study to focus on darker-skin tones for chronic wound segmentation using models trained only on wound images exhibiting lighter skin. Diabetes is highly prevalent in countries where patients have darker skin tones, highlighting the need for a greater focus on such cases. Additionally, we conduct the largest qualitative study to date for chronic wound segmentation.

☆ LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decoding

Doohyuk Jang, Sihwan Park, June Yong Yang, Yeonsung Jung, Jihun Yun, Souvik Kundu, Sung-Yub Kim, Eunho Yang

Auto-Regressive (AR) models have recently gained prominence in image generation, often matching or even surpassing the performance of diffusion models. However, one major limitation of AR models is their sequential nature, which processes tokens one at a time, slowing down generation compared to models like GANs or diffusion-based methods that operate more efficiently. While speculative decoding has proven effective for accelerating LLMs by generating multiple tokens in a single forward, its application in visual AR models remains largely unexplored. In this work, we identify a challenge in this setting, which we term \textit{token selection ambiguity}, wherein visual AR models frequently assign uniformly low probabilities to tokens, hampering the performance of speculative decoding. To overcome this challenge, we propose a relaxed acceptance condition referred to as LANTERN that leverages the interchangeability of tokens in latent space. This relaxation restores the effectiveness of speculative decoding in visual AR models by enabling more flexible use of candidate tokens that would otherwise be prematurely rejected. Furthermore, by incorporating a total variation distance bound, we ensure that these speed gains are achieved without significantly compromising image quality or semantic coherence. Experimental results demonstrate the efficacy of our method in providing a substantial speed-up over speculative decoding. In specific, compared to a na\"ive application of the state-of-the-art speculative decoding, LANTERN increases speed-ups by $\mathbf{1.75}\times$ and $\mathbf{1.76}\times$, as compared to greedy decoding and random sampling, respectively, when applied to LlamaGen, a contemporary visual AR model.

☆ Audio-Agent: Leveraging LLMs For Audio Generation, Editing and Composition

Zixuan Wang, Yu-Wing Tai, Chi-Keung Tang

We introduce Audio-Agent, a multimodal framework for audio generation, editing and composition based on text or video inputs. Conventional approaches for text-to-audio (TTA) tasks often make single-pass inferences from text descriptions. While straightforward, this design struggles to produce high-quality audio when given complex text conditions. In our method, we utilize a pre-trained TTA diffusion network as the audio generation agent to work in tandem with GPT-4, which decomposes the text condition into atomic, specific instructions, and calls the agent for audio generation. Consequently, Audio-Agent generates high-quality audio that is closely aligned with the provided text or video while also supporting variable-length generation. For video-to-audio (VTA) tasks, most existing methods require training a timestamp detector to synchronize video events with generated audio, a process that can be tedious and time-consuming. We propose a simpler approach by fine-tuning a pre-trained Large Language Model (LLM), e.g., Gemma2-2B-it, to obtain both semantic and temporal conditions to bridge video and audio modality. Thus our framework provides a comprehensive solution for both TTA and VTA tasks without substantial computational overhead in training.

☆ An X-Ray Is Worth 15 Features: Sparse Autoencoders for Interpretable Radiology Report Generation

Ahmed Abdulaal, Hugo Fry, Nina Montaña-Brown, Ayodeji Ijishakin, Jack Gao, Stephanie Hyland, Daniel C. Alexander, Daniel C. Castro

Radiological services are experiencing unprecedented demand, leading to increased interest in automating radiology report generation. Existing Vision-Language Models (VLMs) suffer from hallucinations, lack interpretability, and require expensive fine-tuning. We introduce SAE-Rad, which uses sparse autoencoders (SAEs) to decompose latent representations from a pre-trained vision transformer into human-interpretable features. Our hybrid architecture combines state-of-the-art SAE advancements, achieving accurate latent reconstructions while maintaining sparsity. Using an off-the-shelf language model, we distil ground-truth reports into radiological descriptions for each SAE feature, which we then compile into a full report for each image, eliminating the need for fine-tuning large models for this task. To the best of our knowledge, SAE-Rad represents the first instance of using mechanistic interpretability techniques explicitly for a downstream multi-modal reasoning task. On the MIMIC-CXR dataset, SAE-Rad achieves competitive radiology-specific metrics compared to state-of-the-art models while using significantly fewer computational resources for training. Qualitative analysis reveals that SAE-Rad learns meaningful visual concepts and generates reports aligning closely with expert interpretations. Our results suggest that SAEs can enhance multimodal reasoning in healthcare, providing a more interpretable alternative to existing VLMs.

☆ Comparative Analysis and Ensemble Enhancement of Leading CNN Architectures for Breast Cancer Classification

Gary Murphy, Raghubir Singh

This study introduces a novel and accurate approach to breast cancer classification using histopathology images. It systematically compares leading Convolutional Neural Network (CNN) models across varying image datasets, identifies their optimal hyperparameters, and ranks them based on classification efficacy. To maximize classification accuracy for each model we explore, the effects of data augmentation, alternative fully-connected layers, model training hyperparameter settings, and, the advantages of retraining models versus using pre-trained weights. Our methodology includes several original concepts, including serializing generated datasets to ensure consistent data conditions across training runs and significantly reducing training duration. Combined with automated curation of results, this enabled the exploration of over 2,000 training permutations -- such a comprehensive comparison is as yet unprecedented. Our findings establish the settings required to achieve exceptional classification accuracy for standalone CNN models and rank them by model efficacy. Based on these results, we propose ensemble architectures that stack three high-performing standalone CNN models together with diverse classifiers, resulting in improved classification accuracy. The ability to systematically run so many model permutations to get the best outcomes gives rise to very high quality results, including 99.75% for BreakHis x40 and BreakHis x200 and 95.18% for the Bach datasets when split into train, validation and test datasets. The Bach Online blind challenge, yielded 89% using this approach. Whilst this study is based on breast cancer histopathology image datasets, the methodology is equally applicable to other medical image datasets.

☆ EmojiHeroVR: A Study on Facial Expression Recognition under Partial Occlusion from Head-Mounted Displays

Thorben Ortmann, Qi Wang, Larissa Putzar

Emotion recognition promotes the evaluation and enhancement of Virtual Reality (VR) experiences by providing emotional feedback and enabling advanced personalization. However, facial expressions are rarely used to recognize users' emotions, as Head-Mounted Displays (HMDs) occlude the upper half of the face. To address this issue, we conducted a study with 37 participants who played our novel affective VR game EmojiHeroVR. The collected database, EmoHeVRDB (EmojiHeroVR Database), includes 3,556 labeled facial images of 1,778 reenacted emotions. For each labeled image, we also provide 29 additional frames recorded directly before and after the labeled image to facilitate dynamic Facial Expression Recognition (FER). Additionally, EmoHeVRDB includes data on the activations of 63 facial expressions captured via the Meta Quest Pro VR headset for each frame. Leveraging our database, we conducted a baseline evaluation on the static FER classification task with six basic emotions and neutral using the EfficientNet-B0 architecture. The best model achieved an accuracy of 69.84% on the test set, indicating that FER under HMD occlusion is feasible but significantly more challenging than conventional FER.

☆ Does SpatioTemporal information benefit Two video summarization benchmarks? ECAI 2024

Aashutosh Ganesh, Mirela Popa, Daan Odijk, Nava Tintarev

An important aspect of summarizing videos is understanding the temporal context behind each part of the video to grasp what is and is not important. Video summarization models have in recent years modeled spatio-temporal relationships to represent this information. These models achieved state-of-the-art correlation scores on important benchmark datasets. However, what has not been reviewed is whether spatio-temporal relationships are even required to achieve state-of-the-art results. Previous work in activity recognition has found biases, by prioritizing static cues such as scenes or objects, over motion information. In this paper we inquire if similar spurious relationships might influence the task of video summarization. To do so, we analyse the role that temporal information plays on existing benchmark datasets. We first estimate a baseline with temporally invariant models to see how well such models rank on benchmark datasets (TVSum and SumMe). We then disrupt the temporal order of the videos to investigate the impact it has on existing state-of-the-art models. One of our findings is that the temporally invariant models achieve competitive correlation scores that are close to the human baselines on the TVSum dataset. We also demonstrate that existing models are not affected by temporal perturbations. Furthermore, with certain disruption strategies that shuffle fixed time segments, we can actually improve their correlation scores. With these results, we find that spatio-temporal relationship play a minor role and we raise the question whether these benchmarks adequately model the task of video summarization. Code available at: https://github.com/AashGan/TemporalPerturbSum

comment: Accepted for presentation at AEQUITAS workshop, Co-located with ECAI 2024

☆ Visual-O1: Understanding Ambiguous Instructions via Multi-modal Multi-turn Chain-of-thoughts Reasoning

Minheng Ni, Yutao Fan, Lei Zhang, Wangmeng Zuo

As large-scale models evolve, language instructions are increasingly utilized in multi-modal tasks. Due to human language habits, these instructions often contain ambiguities in real-world scenarios, necessitating the integration of visual context or common sense for accurate interpretation. However, even highly intelligent large models exhibit significant performance limitations on ambiguous instructions, where weak reasoning abilities of disambiguation can lead to catastrophic errors. To address this issue, this paper proposes Visual-O1, a multi-modal multi-turn chain-of-thought reasoning framework. It simulates human multi-modal multi-turn reasoning, providing instantial experience for highly intelligent models or empirical experience for generally intelligent models to understand ambiguous instructions. Unlike traditional methods that require models to possess high intelligence to understand long texts or perform lengthy complex reasoning, our framework does not significantly increase computational overhead and is more general and effective, even for generally intelligent models. Experiments show that our method not only significantly enhances the performance of models of different intelligence levels on ambiguous instructions but also improves their performance on general datasets. Our work highlights the potential of artificial intelligence to work like humans in real-world scenarios with uncertainty and ambiguity. We will release our data and code.

☆ Lost in Tracking: Uncertainty-guided Cardiac Cine MRI Segmentation at Right Ventricle Base

Yidong Zhao, Yi Zhang, Orlando Simonetti, Yuchi Han, Qian Tao

Accurate biventricular segmentation of cardiac magnetic resonance (CMR) cine images is essential for the clinical evaluation of heart function. However, compared to left ventricle (LV), right ventricle (RV) segmentation is still more challenging and less reproducible. Degenerate performance frequently occurs at the RV base, where the in-plane anatomical structures are complex (with atria, valve, and aorta) and vary due to the strong interplanar motion. In this work, we propose to address the currently unsolved issues in CMR segmentation, specifically at the RV base, with two strategies: first, we complemented the public resource by reannotating the RV base in the ACDC dataset, with refined delineation of the right ventricle outflow tract (RVOT), under the guidance of an expert cardiologist. Second, we proposed a novel dual encoder U-Net architecture that leverages temporal incoherence to inform the segmentation when interplanar motions occur. The inter-planar motion is characterized by loss-of-tracking, via Bayesian uncertainty of a motion-tracking model. Our experiments showed that our method significantly improved RV base segmentation taking into account temporal incoherence. Furthermore, we investigated the reproducibility of deep learning-based segmentation and showed that the combination of consistent annotation and loss of tracking could enhance the reproducibility of RV segmentation, potentially facilitating a large number of clinical studies focusing on RV.

☆ Quo Vadis, Motion Generation? From Large Language Models to Large Motion Models

Ye Wang, Sipeng Zheng, Bin Cao, Qianshan Wei, Qin Jin, Zongqing Lu

Inspired by the recent success of LLMs, the field of human motion understanding has increasingly shifted towards the development of large motion models. Despite some progress, current state-of-the-art works remain far from achieving truly generalist models, largely due to the lack of large-scale, high-quality motion data. To address this, we present MotionBase, the first million-level motion generation benchmark, offering 15 times the data volume of the previous largest dataset, and featuring multimodal data with hierarchically detailed text descriptions. By leveraging this vast dataset, our large motion model demonstrates strong performance across a broad range of motions, including unseen ones. Through systematic investigation, we underscore the importance of scaling both data and model size, with synthetic data and pseudo labels playing a crucial role in mitigating data acquisition costs. Moreover, our research reveals the limitations of existing evaluation metrics, particularly in handling out-of-domain text instructions -- an issue that has long been overlooked. In addition to these, we introduce a novel 2D lookup-free approach for motion tokenization, which preserves motion information and expands codebook capacity, further enhancing the representative ability of large motion models. The release of MotionBase and the insights gained from this study are expected to pave the way for the development of more powerful and versatile motion generation models.

☆ SELU: Self-Learning Embodied MLLMs in Unknown Environments

Boyu Li, Haobin Jiang, Ziluo Ding, Xinrun Xu, Haoran Li, Dongbin Zhao, Zongqing Lu

Recently, multimodal large language models (MLLMs) have demonstrated strong visual understanding and decision-making capabilities, enabling the exploration of autonomously improving MLLMs in unknown environments. However, external feedback like human or environmental feedback is not always available. To address this challenge, existing methods primarily focus on enhancing the decision-making capabilities of MLLMs through voting and scoring mechanisms, while little effort has been paid to improving the environmental comprehension of MLLMs in unknown environments. To fully unleash the self-learning potential of MLLMs, we propose a novel actor-critic self-learning paradigm, dubbed SELU, inspired by the actor-critic paradigm in reinforcement learning. The critic employs self-asking and hindsight relabeling to extract knowledge from interaction trajectories collected by the actor, thereby augmenting its environmental comprehension. Simultaneously, the actor is improved by the self-feedback provided by the critic, enhancing its decision-making. We evaluate our method in the AI2-THOR and VirtualHome environments, and SELU achieves critic improvements of approximately 28% and 30%, and actor improvements of about 20% and 24% via self-learning.

☆ Action Selection Learning for Multi-label Multi-view Action Recognition

Trung Thanh Nguyen, Yasutomo Kawanishi, Takahiro Komamizu, Ichiro Ide

Multi-label multi-view action recognition aims to recognize multiple concurrent or sequential actions from untrimmed videos captured by multiple cameras. Existing work has focused on multi-view action recognition in a narrow area with strong labels available, where the onset and offset of each action are labeled at the frame-level. This study focuses on real-world scenarios where cameras are distributed to capture a wide-range area with only weak labels available at the video-level. We propose the method named MultiASL (Multi-view Action Selection Learning), which leverages action selection learning to enhance view fusion by selecting the most useful information from different viewpoints. The proposed method includes a Multi-view Spatial-Temporal Transformer video encoder to extract spatial and temporal features from multi-viewpoint videos. Action Selection Learning is employed at the frame-level, using pseudo ground-truth obtained from weak labels at the video-level, to identify the most relevant frames for action recognition. Experiments in a real-world office environment using the MM-Office dataset demonstrate the superior performance of the proposed method compared to existing methods.

comment: ACM Multimedia Asia 2024

☆ Grounded-VideoLLM: Sharpening Fine-grained Temporal Grounding in Video Large Language Models

Haibo Wang, Zhiyang Xu, Yu Cheng, Shizhe Diao, Yufan Zhou, Yixin Cao, Qifan Wang, Weifeng Ge, Lifu Huang

Video Large Language Models (Video-LLMs) have demonstrated remarkable capabilities in coarse-grained video understanding, however, they struggle with fine-grained temporal grounding. In this paper, we introduce Grounded-VideoLLM, a novel Video-LLM adept at perceiving and reasoning over specific video moments in a fine-grained manner. We identify that current Video-LLMs have limitations for fine-grained video understanding since they lack effective temporal modeling and timestamp representation. In light of this, we sharpen our model by incorporating (1) an additional temporal stream to encode the relationships between frames and (2) discrete temporal tokens enriched with specific time knowledge to represent timestamps. To optimize the training of Grounded-VideoLLM, we employ a multi-stage training scheme, beginning with simple video-captioning tasks and progressively introducing video temporal grounding tasks of increasing complexity. To further enhance Grounded-VideoLLM's temporal reasoning capability, we also curate a grounded VideoQA dataset by an automatic annotation pipeline. Extensive experiments demonstrate that Grounded-VideoLLM not only excels in fine-grained grounding tasks such as temporal sentence grounding, dense video captioning, and grounded VideoQA, but also shows great potential as a versatile video assistant for general video understanding.

☆ Semantic Segmentation Based Quality Control of Histopathology Whole Slide Images

Abhijeet Patil, Garima Jain, Harsh Diwakar, Jay Sawant, Tripti Bameta, Swapnil Rane, Amit Sethi

We developed a software pipeline for quality control (QC) of histopathology whole slide images (WSIs) that segments various regions, such as blurs of different levels, tissue regions, tissue folds, and pen marks. Given the necessity and increasing availability of GPUs for processing WSIs, the proposed pipeline comprises multiple lightweight deep learning models to strike a balance between accuracy and speed. The pipeline was evaluated in all TCGAs, which is the largest publicly available WSI dataset containing more than 11,000 histopathological images from 28 organs. It was compared to a previous work, which was not based on deep learning, and it showed consistent improvement in segmentation results across organs. To minimize annotation effort for tissue and blur segmentation, annotated images were automatically prepared by mosaicking patches (sub-images) from various WSIs whose labels were identified using a patch classification tool HistoROI. Due to the generality of our trained QC pipeline and its extensive testing the potential impact of this work is broad. It can be used for automated pre-processing any WSI cohort to enhance the accuracy and reliability of large-scale histopathology image analysis for both research and clinical use. We have made the trained models, training scripts, training data, and inference results publicly available at https://github.com/abhijeetptl5/wsisegqc, which should enable the research community to use the pipeline right out of the box or further customize it to new datasets and applications in the future.

comment: 14 pages, 8 figures

☆ Sm: enhanced localization in Multiple Instance Learning for medical imaging classification NeurIPS 2024

Francisco M. Castro-Macías, Pablo Morales-Álvarez, Yunan Wu, Rafael Molina, Aggelos K. Katsaggelos

Multiple Instance Learning (MIL) is widely used in medical imaging classification to reduce the labeling effort. While only bag labels are available for training, one typically seeks predictions at both bag and instance levels (classification and localization tasks, respectively). Early MIL methods treated the instances in a bag independently. Recent methods account for global and local dependencies among instances. Although they have yielded excellent results in classification, their performance in terms of localization is comparatively limited. We argue that these models have been designed to target the classification task, while implications at the instance level have not been deeply investigated. Motivated by a simple observation -- that neighboring instances are likely to have the same label -- we propose a novel, principled, and flexible mechanism to model local dependencies. It can be used alone or combined with any mechanism to model global dependencies (e.g., transformers). A thorough empirical validation shows that our module leads to state-of-the-art performance in localization while being competitive or superior in classification. Our code is at https://github.com/Franblueee/SmMIL.

comment: 24 pages, 14 figures, 2024 Conference on Neural Information Processing Systems (NeurIPS 2024)

☆ 3D Segmentation of Neuronal Nuclei and Cell-Type Identification using Multi-channel Information

Antonio LaTorre, Lidia Alonso-Nanclares, José María Peña, Javier De Felipe

Background Analyzing images to accurately estimate the number of different cell types in the brain using automatic methods is a major objective in neuroscience. The automatic and selective detection and segmentation of neurons would be an important step in neuroanatomical studies. New method We present a method to improve the 3D reconstruction of neuronal nuclei that allows their segmentation, excluding the nuclei of non-neuronal cell types. Results We have tested the algorithm on stacks of images from rat neocortex, in a complex scenario (large stacks of images, uneven staining, and three different channels to visualize different cellular markers). It was able to provide a good identification ratio of neuronal nuclei and a 3D segmentation. Comparison with Existing Methods: Many automatic tools are in fact currently available, but different methods yield different cell count estimations, even in the same brain regions, due to differences in the labeling and imaging techniques, as well as in the algorithms used to detect cells. Moreover, some of the available automated software methods have provided estimations of cell numbers that have been reported to be inaccurate or inconsistent after evaluation by neuroanatomists. Conclusions It is critical to have a tool for automatic segmentation that allows discrimination between neurons, glial cells and perivascular cells. It would greatly speed up a task that is currently performed manually and would allow the cell counting to be systematic, avoiding human bias. Furthermore, the resulting 3D reconstructions of different cell types can be used to generate models of the spatial distribution of cells.

☆ ScriptViz: A Visualization Tool to Aid Scriptwriting based on a Large Movie Database

Anyi Rao, Jean-Peïc Chou, Maneesh Agrawala

Scriptwriters usually rely on their mental visualization to create a vivid story by using their imagination to see, feel, and experience the scenes they are writing. Besides mental visualization, they often refer to existing images or scenes in movies and analyze the visual elements to create a certain mood or atmosphere. In this paper, we develop ScriptViz to provide external visualization based on a large movie database for the screenwriting process. It retrieves reference visuals on the fly based on scripts' text and dialogue from a large movie database. The tool provides two types of control on visual elements that enable writers to 1) see exactly what they want with fixed visual elements and 2) see variances in uncertain elements. User evaluation among 15 scriptwriters shows that ScriptViz is able to present scriptwriters with consistent yet diverse visual possibilities, aligning closely with their scripts and helping their creation.

comment: Accepted in the 37th Annual ACM Symposium on User Interface Software and Technology (UIST'24). Webpage: https://virtualfilmstudio.github.io/projects/scriptviz

☆ Tuning Timestep-Distilled Diffusion Model Using Pairwise Sample Optimization

Zichen Miao, Zhengyuan Yang, Kevin Lin, Ze Wang, Zicheng Liu, Lijuan Wang, Qiang Qiu

Recent advancements in timestep-distilled diffusion models have enabled high-quality image generation that rivals non-distilled multi-step models, but with significantly fewer inference steps. While such models are attractive for applications due to the low inference cost and latency, fine-tuning them with a naive diffusion objective would result in degraded and blurry outputs. An intuitive alternative is to repeat the diffusion distillation process with a fine-tuned teacher model, which produces good results but is cumbersome and computationally intensive; the distillation training usually requires magnitude higher of training compute compared to fine-tuning for specific image styles. In this paper, we present an algorithm named pairwise sample optimization (PSO), which enables the direct fine-tuning of an arbitrary timestep-distilled diffusion model. PSO introduces additional reference images sampled from the current time-step distilled model, and increases the relative likelihood margin between the training images and reference images. This enables the model to retain its few-step generation ability, while allowing for fine-tuning of its output distribution. We also demonstrate that PSO is a generalized formulation which can be flexibly extended to both offline-sampled and online-sampled pairwise data, covering various popular objectives for diffusion model preference optimization. We evaluate PSO in both preference optimization and other fine-tuning tasks, including style transfer and concept customization. We show that PSO can directly adapt distilled models to human-preferred generation with both offline and online-generated pairwise preference image data. PSO also demonstrates effectiveness in style transfer and concept customization by directly tuning timestep-distilled diffusion models.

☆ Generalizable Prompt Tuning for Vision-Language Models

Qian Zhang

Prompt tuning for vision-language models such as CLIP involves optimizing the text prompts used to generate image-text pairs for specific downstream tasks. While hand-crafted or template-based prompts are generally applicable to a wider range of unseen classes, they tend to perform poorly in downstream tasks (i.e., seen classes). Learnable soft prompts, on the other hand, often perform well in downstream tasks but lack generalizability. Additionally, prior research has predominantly concentrated on the textual modality, with very few studies attempting to explore the prompt's generalization potential from the visual modality. Keeping these limitations in mind, we investigate how to prompt tuning to obtain both a competitive downstream performance and generalization. The study shows that by treating soft and hand-crafted prompts as dual views of the textual modality, and maximizing their mutual information, we can better ensemble task-specific and general semantic information. Moreover, to generate more expressive prompts, the study introduces a class-wise augmentation from the visual modality, resulting in significant robustness to a wider range of unseen classes. Extensive evaluations on several benchmarks report that the proposed approach achieves competitive results in terms of both task-specific performance and general abilities.

☆ Looking into Concept Explanation Methods for Diabetic Retinopathy Classification

Andrea M. Storås, Josefine V. Sundgaard

Diabetic retinopathy is a common complication of diabetes, and monitoring the progression of retinal abnormalities using fundus imaging is crucial. Because the images must be interpreted by a medical expert, it is infeasible to screen all individuals with diabetes for diabetic retinopathy. Deep learning has shown impressive results for automatic analysis and grading of fundus images. One drawback is, however, the lack of interpretability, which hampers the implementation of such systems in the clinic. Explainable artificial intelligence methods can be applied to explain the deep neural networks. Explanations based on concepts have shown to be intuitive for humans to understand, but have not yet been explored in detail for diabetic retinopathy grading. This work investigates and compares two concept-based explanation techniques for explaining deep neural networks developed for automatic diagnosis of diabetic retinopathy: Quantitative Testing with Concept Activation Vectors and Concept Bottleneck Models. We found that both methods have strengths and weaknesses, and choice of method should take the available data and the end user's preferences into account.

comment: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://melba-journal.org/2024:021

☆ Autonomous Character-Scene Interaction Synthesis from Text Instruction

Nan Jiang, Zimo He, Zi Wang, Hongjie Li, Yixin Chen, Siyuan Huang, Yixin Zhu

Synthesizing human motions in 3D environments, particularly those with complex activities such as locomotion, hand-reaching, and human-object interaction, presents substantial demands for user-defined waypoints and stage transitions. These requirements pose challenges for current models, leading to a notable gap in automating the animation of characters from simple human inputs. This paper addresses this challenge by introducing a comprehensive framework for synthesizing multi-stage scene-aware interaction motions directly from a single text instruction and goal location. Our approach employs an auto-regressive diffusion model to synthesize the next motion segment, along with an autonomous scheduler predicting the transition for each action stage. To ensure that the synthesized motions are seamlessly integrated within the environment, we propose a scene representation that considers the local perception both at the start and the goal location. We further enhance the coherence of the generated motion by integrating frame embeddings with language input. Additionally, to support model training, we present a comprehensive motion-captured dataset comprising 16 hours of motion sequences in 120 indoor scenes covering 40 types of motions, each annotated with precise language descriptions. Experimental results demonstrate the efficacy of our method in generating high-quality, multi-stage motions closely aligned with environmental and textual conditions.

☆ Investigating and Mitigating Object Hallucinations in Pretrained Vision-Language (CLIP) Models EMNLP 2024

Yufang Liu, Tao Ji, Changzhi Sun, Yuanbin Wu, Aimin Zhou

Large Vision-Language Models (LVLMs) have achieved impressive performance, yet research has pointed out a serious issue with object hallucinations within these models. However, there is no clear conclusion as to which part of the model these hallucinations originate from. In this paper, we present an in-depth investigation into the object hallucination problem specifically within the CLIP model, which serves as the backbone for many state-of-the-art vision-language systems. We unveil that even in isolation, the CLIP model is prone to object hallucinations, suggesting that the hallucination problem is not solely due to the interaction between vision and language modalities. To address this, we propose a counterfactual data augmentation method by creating negative samples with a variety of hallucination issues. We demonstrate that our method can effectively mitigate object hallucinations for CLIP model, and we show the the enhanced model can be employed as a visual encoder, effectively alleviating the object hallucination issue in LVLMs.

comment: EMNLP 2024

☆ HRVMamba: High-Resolution Visual State Space Model for Dense Prediction

Hao Zhang, Yongqiang Ma, Wenqi Shao, Ping Luo, Nanning Zheng, Kaipeng Zhang

Recently, State Space Models (SSMs) with efficient hardware-aware designs, i.e., Mamba, have demonstrated significant potential in computer vision tasks due to their linear computational complexity with respect to token length and their global receptive field. However, Mamba's performance on dense prediction tasks, including human pose estimation and semantic segmentation, has been constrained by three key challenges: insufficient inductive bias, long-range forgetting, and low-resolution output representation. To address these challenges, we introduce the Dynamic Visual State Space (DVSS) block, which utilizes multi-scale convolutional kernels to extract local features across different scales and enhance inductive bias, and employs deformable convolution to mitigate the long-range forgetting problem while enabling adaptive spatial aggregation based on input and task-specific information. By leveraging the multi-resolution parallel design proposed in HRNet, we introduce High-Resolution Visual State Space Model (HRVMamba) based on the DVSS block, which preserves high-resolution representations throughout the entire process while promoting effective multi-scale feature learning. Extensive experiments highlight HRVMamba's impressive performance on dense prediction tasks, achieving competitive results against existing benchmark models without bells and whistles. Code is available at https://github.com/zhanghao5201/HRVMamba.

☆ Selective Transformer for Hyperspectral Image Classification

Yichu Xu, Di Wang, Lefei Zhang, Liangpei Zhang

☆ Redefining Temporal Modeling in Video Diffusion: The Vectorized Timestep Approach

Yaofang Liu, Yumeng Ren, Xiaodong Cun, Aitor Artola, Yang Liu, Tieyong Zeng, Raymond H. Chan, Jean-michel Morel

Diffusion models have revolutionized image generation, and their extension to video generation has shown promise. However, current video diffusion models~(VDMs) rely on a scalar timestep variable applied at the clip level, which limits their ability to model complex temporal dependencies needed for various tasks like image-to-video generation. To address this limitation, we propose a frame-aware video diffusion model~(FVDM), which introduces a novel vectorized timestep variable~(VTV). Unlike conventional VDMs, our approach allows each frame to follow an independent noise schedule, enhancing the model's capacity to capture fine-grained temporal dependencies. FVDM's flexibility is demonstrated across multiple tasks, including standard video generation, image-to-video generation, video interpolation, and long video synthesis. Through a diverse set of VTV configurations, we achieve superior quality in generated videos, overcoming challenges such as catastrophic forgetting during fine-tuning and limited generalizability in zero-shot methods.Our empirical evaluations show that FVDM outperforms state-of-the-art methods in video generation quality, while also excelling in extended tasks. By addressing fundamental shortcomings in existing VDMs, FVDM sets a new paradigm in video synthesis, offering a robust framework with significant implications for generative modeling and multimedia applications.

comment: Code at https://github.com/Yaofang-Liu/FVDM

☆ Bridging the Gap between Text, Audio, Image, and Any Sequence: A Novel Approach using Gloss-based Annotation

Sen Fang, Yalin Feng, Sizhou Chen, Xiaofeng Zhang, Teik Toe Teoh

This paper presents an innovative approach called BGTAI to simplify multimodal understanding by utilizing gloss-based annotation as an intermediate step in aligning Text and Audio with Images. While the dynamic temporal factors in textual and audio inputs contain various predicate adjectives that influence the meaning of the entire sentence, images, on the other hand, present static scenes. By representing text and audio as gloss notations that omit complex semantic nuances, a better alignment with images can potentially be achieved. This study explores the feasibility of this idea, specifically, we first propose the first Langue2Gloss model and then integrate it into the multimodal model UniBriVL for joint training. To strengthen the adaptability of gloss with text/audio and overcome the efficiency and instability issues in multimodal training, we propose a DS-Net (Data-Pair Selection Network), an Result Filter module, and a novel SP-Loss function. Our approach outperforms previous multimodal models in the main experiments, demonstrating its efficacy in enhancing multimodal representations and improving compatibility among text, audio, visual, and any sequence modalities.

☆ ECHOPulse: ECG controlled echocardio-grams video generation

Yiwei Li, Sekeun Kim, Zihao Wu, Hanqi Jiang, Yi Pan, Pengfei Jin, Sifan Song, Yucheng Shi, Tianze Yang, Tianming Liu, Quanzheng Li, Xiang Li

Echocardiography (ECHO) is essential for cardiac assessments, but its video quality and interpretation heavily relies on manual expertise, leading to inconsistent results from clinical and portable devices. ECHO video generation offers a solution by improving automated monitoring through synthetic data and generating high-quality videos from routine health data. However, existing models often face high computational costs, slow inference, and rely on complex conditional prompts that require experts' annotations. To address these challenges, we propose ECHOPULSE, an ECG-conditioned ECHO video generation model. ECHOPULSE introduces two key advancements: (1) it accelerates ECHO video generation by leveraging VQ-VAE tokenization and masked visual token modeling for fast decoding, and (2) it conditions on readily accessible ECG signals, which are highly coherent with ECHO videos, bypassing complex conditional prompts. To the best of our knowledge, this is the first work to use time-series prompts like ECG signals for ECHO video generation. ECHOPULSE not only enables controllable synthetic ECHO data generation but also provides updated cardiac function information for disease monitoring and prediction beyond ECG alone. Evaluations on three public and private datasets demonstrate state-of-the-art performance in ECHO video generation across both qualitative and quantitative measures. Additionally, ECHOPULSE can be easily generalized to other modality generation tasks, such as cardiac MRI, fMRI, and 3D CT generation. Demo can seen from \url{https://github.com/levyisthebest/ECHOPulse_Prelease}.

☆ ARB-LLM: Alternating Refined Binarizations for Large Language Models

Zhiteng Li, Xianglong Yan, Tianao Zhang, Haotong Qin, Dong Xie, Jiang Tian, zhongchao shi, Linghe Kong, Yulun Zhang, Xiaokang Yang

Large Language Models (LLMs) have greatly pushed forward advancements in natural language processing, yet their high memory and computational demands hinder practical deployment. Binarization, as an effective compression technique, can shrink model weights to just 1 bit, significantly reducing the high demands on computation and memory. However, current binarization methods struggle to narrow the distribution gap between binarized and full-precision weights, while also overlooking the column deviation in LLM weight distribution. To tackle these issues, we propose ARB-LLM, a novel 1-bit post-training quantization (PTQ) technique tailored for LLMs. To narrow the distribution shift between binarized and full-precision weights, we first design an alternating refined binarization (ARB) algorithm to progressively update the binarization parameters, which significantly reduces the quantization error. Moreover, considering the pivot role of calibration data and the column deviation in LLM weights, we further extend ARB to ARB-X and ARB-RC. In addition, we refine the weight partition strategy with column-group bitmap (CGB), which further enhance performance. Equipping ARB-X and ARB-RC with CGB, we obtain ARB-LLM$_\text{X}$ and ARB-LLM$_\text{RC}$ respectively, which significantly outperform state-of-the-art (SOTA) binarization methods for LLMs. As a binary PTQ method, our ARB-LLM$_\text{RC}$ is the first to surpass FP16 models of the same size. The code and models will be available at https://github.com/ZHITENGLI/ARB-LLM.

comment: The code and models will be available at https://github.com/ZHITENGLI/ARB-LLM

☆ MBDS: A Multi-Body Dynamics Simulation Dataset for Graph Networks Simulators

Sheng Yang, Fengge Wu, Junsuo Zhao

Modeling the structure and events of the physical world constitutes a fundamental objective of neural networks. Among the diverse approaches, Graph Network Simulators (GNS) have emerged as the leading method for modeling physical phenomena, owing to their low computational cost and high accuracy. The datasets employed for training and evaluating physical simulation techniques are typically generated by researchers themselves, often resulting in limited data volume and quality. Consequently, this poses challenges in accurately assessing the performance of these methods. In response to this, we have constructed a high-quality physical simulation dataset encompassing 1D, 2D, and 3D scenes, along with more trajectories and time-steps compared to existing datasets. Furthermore, our work distinguishes itself by developing eight complete scenes, significantly enhancing the dataset's comprehensiveness. A key feature of our dataset is the inclusion of precise multi-body dynamics, facilitating a more realistic simulation of the physical world. Utilizing our high-quality dataset, we conducted a systematic evaluation of various existing GNS methods. Our dataset is accessible for download at https://github.com/Sherlocktein/MBDS, offering a valuable resource for researchers to enhance the training and evaluation of their methodologies.

☆ Mamba in Vision: A Comprehensive Survey of Techniques and Applications

Md Maklachur Rahman, Abdullah Aman Tutul, Ankur Nath, Lamyanba Laishram, Soon Ki Jung, Tracy Hammond

Mamba is emerging as a novel approach to overcome the challenges faced by Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) in computer vision. While CNNs excel at extracting local features, they often struggle to capture long-range dependencies without complex architectural modifications. In contrast, ViTs effectively model global relationships but suffer from high computational costs due to the quadratic complexity of their self-attention mechanisms. Mamba addresses these limitations by leveraging Selective Structured State Space Models to effectively capture long-range dependencies with linear computational complexity. This survey analyzes the unique contributions, computational benefits, and applications of Mamba models while also identifying challenges and potential future research directions. We provide a foundational resource for advancing the understanding and growth of Mamba models in computer vision. An overview of this work is available at https://github.com/maklachur/Mamba-in-Computer-Vision.

comment: Under Review

☆ Combing Text-based and Drag-based Editing for Precise and Flexible Image Editing

Ziqi Jiang, Zhen Wang, Long Chen

Precise and flexible image editing remains a fundamental challenge in computer vision. Based on the modified areas, most editing methods can be divided into two main types: global editing and local editing. In this paper, we choose the two most common editing approaches (ie text-based editing and drag-based editing) and analyze their drawbacks. Specifically, text-based methods often fail to describe the desired modifications precisely, while drag-based methods suffer from ambiguity. To address these issues, we proposed \textbf{CLIPDrag}, a novel image editing method that is the first to combine text and drag signals for precise and ambiguity-free manipulations on diffusion models. To fully leverage these two signals, we treat text signals as global guidance and drag points as local information. Then we introduce a novel global-local motion supervision method to integrate text signals into existing drag-based methods by adapting a pre-trained language-vision model like CLIP. Furthermore, we also address the problem of slow convergence in CLIPDrag by presenting a fast point-tracking method that enforces drag points moving toward correct directions. Extensive experiments demonstrate that CLIPDrag outperforms existing single drag-based methods or text-based methods.

comment: 12 pages, 9 figures

☆ Generative Edge Detection with Stable Diffusion

Caixia Zhou, Yaping Huang, Mochu Xiang, Jiahui Ren, Haibin Ling, Jing Zhang

Edge detection is typically viewed as a pixel-level classification problem mainly addressed by discriminative methods. Recently, generative edge detection methods, especially diffusion model based solutions, are initialized in the edge detection task. Despite great potential, the retraining of task-specific designed modules and multi-step denoising inference limits their broader applications. Upon closer investigation, we speculate that part of the reason is the under-exploration of the rich discriminative information encoded in extensively pre-trained large models (\eg, stable diffusion models). Thus motivated, we propose a novel approach, named Generative Edge Detector (GED), by fully utilizing the potential of the pre-trained stable diffusion model. Our model can be trained and inferred efficiently without specific network design due to the rich high-level and low-level prior knowledge empowered by the pre-trained stable diffusion. Specifically, we propose to finetune the denoising U-Net and predict latent edge maps directly, by taking the latent image feature maps as input. Additionally, due to the subjectivity and ambiguity of the edges, we also incorporate the granularity of the edges into the denoising U-Net model as one of the conditions to achieve controllable and diverse predictions. Furthermore, we devise a granularity regularization to ensure the relative granularity relationship of the multiple predictions. We conduct extensive experiments on multiple datasets and achieve competitive performance (\eg, 0.870 and 0.880 in terms of ODS and OIS on the BSDS test dataset).

☆ DocKD: Knowledge Distillation from LLMs for Open-World Document Understanding Models EMNLP 2024

Sungnyun Kim, Haofu Liao, Srikar Appalaraju, Peng Tang, Zhuowen Tu, Ravi Kumar Satzoda, R. Manmatha, Vijay Mahadevan, Stefano Soatto

Visual document understanding (VDU) is a challenging task that involves understanding documents across various modalities (text and image) and layouts (forms, tables, etc.). This study aims to enhance generalizability of small VDU models by distilling knowledge from LLMs. We identify that directly prompting LLMs often fails to generate informative and useful data. In response, we present a new framework (called DocKD) that enriches the data generation process by integrating external document knowledge. Specifically, we provide an LLM with various document elements like key-value pairs, layouts, and descriptions, to elicit open-ended answers. Our experiments show that DocKD produces high-quality document annotations and surpasses the direct knowledge distillation approach that does not leverage external document knowledge. Moreover, student VDU models trained with solely DocKD-generated data are not only comparable to those trained with human-annotated data on in-domain tasks but also significantly excel them on out-of-domain tasks.

comment: Accepted to EMNLP 2024

☆ DiffKillR: Killing and Recreating Diffeomorphisms for Cell Annotation in Dense Microscopy Images

Chen Liu, Danqi Liao, Alejandro Parada-Mayorga, Alejandro Ribeiro, Marcello DiStasio, Smita Krishnaswamy

The proliferation of digital microscopy images, driven by advances in automated whole slide scanning, presents significant opportunities for biomedical research and clinical diagnostics. However, accurately annotating densely packed information in these images remains a major challenge. To address this, we introduce DiffKillR, a novel framework that reframes cell annotation as the combination of archetype matching and image registration tasks. DiffKillR employs two complementary neural networks: one that learns a diffeomorphism-invariant feature space for robust cell matching and another that computes the precise warping field between cells for annotation mapping. Using a small set of annotated archetypes, DiffKillR efficiently propagates annotations across large microscopy images, reducing the need for extensive manual labeling. More importantly, it is suitable for any type of pixel-level annotation. We will discuss the theoretical properties of DiffKillR and validate it on three microscopy tasks, demonstrating its advantages over existing supervised, semi-supervised, and unsupervised methods.

☆ CLIP-Clique: Graph-based Correspondence Matching Augmented by Vision Language Models for Object-based Global Localization

Shigemichi Matsuzaki, Kazuhito Tanaka, Kazuhiro Shintani

This letter proposes a method of global localization on a map with semantic object landmarks. One of the most promising approaches for localization on object maps is to use semantic graph matching using landmark descriptors calculated from the distribution of surrounding objects. These descriptors are vulnerable to misclassification and partial observations. Moreover, many existing methods rely on inlier extraction using RANSAC, which is stochastic and sensitive to a high outlier rate. To address the former issue, we augment the correspondence matching using Vision Language Models (VLMs). Landmark discriminability is improved by VLM embeddings, which are independent of surrounding objects. In addition, inliers are estimated deterministically using a graph-theoretic approach. We also incorporate pose calculation using the weighted least squares considering correspondence similarity and observation completeness to improve the robustness. We confirmed improvements in matching and pose estimation accuracy through experiments on ScanNet and TUM datasets.

comment: IEEE Robotics and Automation Letters

☆ AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark

Wenhao Chai, Enxin Song, Yilun Du, Chenlin Meng, Vashisht Madhavan, Omer Bar-Tal, Jeng-Neng Hwang, Saining Xie, Christopher D. Manning

Video detailed captioning is a key task which aims to generate comprehensive and coherent textual descriptions of video content, benefiting both video understanding and generation. In this paper, we propose AuroraCap, a video captioner based on a large multimodal model. We follow the simplest architecture design without additional parameters for temporal modeling. To address the overhead caused by lengthy video sequences, we implement the token merging strategy, reducing the number of input visual tokens. Surprisingly, we found that this strategy results in little performance loss. AuroraCap shows superior performance on various video and image captioning benchmarks, for example, obtaining a CIDEr of 88.9 on Flickr30k, beating GPT-4V (55.3) and Gemini-1.5 Pro (82.2). However, existing video caption benchmarks only include simple descriptions, consisting of a few dozen words, which limits research in this field. Therefore, we develop VDC, a video detailed captioning benchmark with over one thousand carefully annotated structured captions. In addition, we propose a new LLM-assisted metric VDCscore for bettering evaluation, which adopts a divide-and-conquer strategy to transform long caption evaluation into multiple short question-answer pairs. With the help of human Elo ranking, our experiments show that this benchmark better correlates with human judgments of video detailed captioning quality.

comment: Code, docs, weight, benchmark and training data are all avaliable at \href{https://rese1f.github.io/aurora-web/}{website}

♻ ☆ Video Instruction Tuning With Synthetic Data

Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, Chunyuan Li

comment: Project page: https://llava-vl.github.io/blog/2024-09-30-llava-video/

♻ ☆ Learning 3D Perception from Others' Predictions

Jinsu Yoo, Zhenyang Feng, Tai-Yu Pan, Yihong Sun, Cheng Perng Phoo, Xiangyu Chen, Mark Campbell, Kilian Q. Weinberger, Bharath Hariharan, Wei-Lun Chao

Accurate 3D object detection in real-world environments requires a huge amount of annotated data with high quality. Acquiring such data is tedious and expensive, and often needs repeated effort when a new sensor is adopted or when the detector is deployed in a new environment. We investigate a new scenario to construct 3D object detectors: learning from the predictions of a nearby unit that is equipped with an accurate detector. For example, when a self-driving car enters a new area, it may learn from other traffic participants whose detectors have been optimized for that area. This setting is label-efficient, sensor-agnostic, and communication-efficient: nearby units only need to share the predictions with the ego agent (e.g., car). Naively using the received predictions as ground-truths to train the detector for the ego car, however, leads to inferior performance. We systematically study the problem and identify viewpoint mismatches and mislocalization (due to synchronization and GPS errors) as the main causes, which unavoidably result in false positives, false negatives, and inaccurate pseudo labels. We propose a distance-based curriculum, first learning from closer units with similar viewpoints and subsequently improving the quality of other units' predictions via self-training. We further demonstrate that an effective pseudo label refinement module can be trained with a handful of annotated data, largely reducing the data quantity necessary to train an object detector. We validate our approach on the recently released real-world collaborative driving dataset, using reference cars' predictions as pseudo labels for the ego car. Extensive experiments including several scenarios (e.g., different sensors, detectors, and domains) demonstrate the effectiveness of our approach toward label-efficient learning of 3D perception from other units' predictions.

♻ ☆ Context and Geometry Aware Voxel Transformer for Semantic Scene Completion NIPS 2024

Zhu Yu, Runmin Zhang, Jiacheng Ying, Junchen Yu, Xiaohai Hu, Lun Luo, Si-Yuan Cao, Hui-Liang Shen

Vision-based Semantic Scene Completion (SSC) has gained much attention due to its widespread applications in various 3D perception tasks. Existing sparse-to-dense approaches typically employ shared context-independent queries across various input images, which fails to capture distinctions among them as the focal regions of different inputs vary and may result in undirected feature aggregation of cross-attention. Additionally, the absence of depth information may lead to points projected onto the image plane sharing the same 2D position or similar sampling points in the feature map, resulting in depth ambiguity. In this paper, we present a novel context and geometry aware voxel transformer. It utilizes a context aware query generator to initialize context-dependent queries tailored to individual input images, effectively capturing their unique characteristics and aggregating information within the region of interest. Furthermore, it extend deformable cross-attention from 2D to 3D pixel space, enabling the differentiation of points with similar image coordinates based on their depth coordinates. Building upon this module, we introduce a neural network named CGFormer to achieve semantic scene completion. Simultaneously, CGFormer leverages multiple 3D representations (i.e., voxel and TPV) to boost the semantic and geometric representation abilities of the transformed 3D volume from both local and global perspectives. Experimental results demonstrate that CGFormer achieves state-of-the-art performance on the SemanticKITTI and SSCBench-KITTI-360 benchmarks, attaining a mIoU of 16.87 and 20.05, as well as an IoU of 45.99 and 48.07, respectively. Remarkably, CGFormer even outperforms approaches employing temporal images as inputs or much larger image backbone networks.

comment: NIPS 2024 Spotlight

♻ ☆ Generalizable Human Gaussians from Single-View Image

Jinnan Chen, Chen Li, Jianfeng Zhang, Lingting Zhu, Buzhen Huang, Hanlin Chen, Gim Hee Lee

In this work, we tackle the task of learning 3D human Gaussians from a single image, focusing on recovering detailed appearance and geometry including unobserved regions. We introduce a single-view generalizable Human Gaussian Model (HGM), which employs a novel generate-then-refine pipeline with the guidance from human body prior and diffusion prior. Our approach uses a ControlNet to refine rendered back-view images from coarse predicted human Gaussians, then uses the refined image along with the input image to reconstruct refined human Gaussians. To mitigate the potential generation of unrealistic human poses and shapes, we incorporate human priors from the SMPL-X model as a dual branch, propagating image features from the SMPL-X volume to the image Gaussians using sparse convolution and attention mechanisms. Given that the initial SMPL-X estimation might be inaccurate, we gradually refine it with our HGM model. We validate our approach on several publicly available datasets. Our method surpasses previous methods in both novel view synthesis and surface reconstruction. Our approach also exhibits strong generalization for cross-dataset evaluation and in-the-wild images.

comment: https://jinnan-chen.github.io/projects/HGM/

♻ ☆ Unleashing the Potential of the Diffusion Model in Few-shot Semantic Segmentation NeurIPS

Muzhi Zhu, Yang Liu, Zekai Luo, Chenchen Jing, Hao Chen, Guangkai Xu, Xinlong Wang, Chunhua Shen

The Diffusion Model has not only garnered noteworthy achievements in the realm of image generation but has also demonstrated its potential as an effective pretraining method utilizing unlabeled data. Drawing from the extensive potential unveiled by the Diffusion Model in both semantic correspondence and open vocabulary segmentation, our work initiates an investigation into employing the Latent Diffusion Model for Few-shot Semantic Segmentation. Recently, inspired by the in-context learning ability of large language models, Few-shot Semantic Segmentation has evolved into In-context Segmentation tasks, morphing into a crucial element in assessing generalist segmentation models. In this context, we concentrate on Few-shot Semantic Segmentation, establishing a solid foundation for the future development of a Diffusion-based generalist model for segmentation. Our initial focus lies in understanding how to facilitate interaction between the query image and the support image, resulting in the proposal of a KV fusion method within the self-attention framework. Subsequently, we delve deeper into optimizing the infusion of information from the support mask and simultaneously re-evaluating how to provide reasonable supervision from the query mask. Based on our analysis, we establish a simple and effective framework named DiffewS, maximally retaining the original Latent Diffusion Model's generative framework and effectively utilizing the pre-training prior. Experimental results demonstrate that our method significantly outperforms the previous SOTA models in multiple settings.

comment: Accepted to Proc. Annual Conference on Neural Information Processing Systems (NeurIPS) 2024

♻ ☆ Decoupling Layout from Glyph in Online Chinese Handwriting Generation

Min-Si Ren, Yan-Ming Zhang, Yi Chen

Text plays a crucial role in the transmission of human civilization, and teaching machines to generate online handwritten text in various styles presents an interesting and significant challenge. However, most prior work has concentrated on generating individual Chinese fonts, leaving {complete text line generation largely unexplored}. In this paper, we identify that text lines can naturally be divided into two components: layout and glyphs. Based on this division, we designed a text line layout generator coupled with a diffusion-based stylized font synthesizer to address this challenge hierarchically. More concretely, the layout generator performs in-context-like learning based on the text content and the provided style references to generate positions for each glyph autoregressively. Meanwhile, the font synthesizer which consists of a character embedding dictionary, a multi-scale calligraphy style encoder, and a 1D U-Net based diffusion denoiser will generate each font on its position while imitating the calligraphy style extracted from the given style references. Qualitative and quantitative experiments on the CASIA-OLHWDB demonstrate that our method is capable of generating structurally correct and indistinguishable imitation samples.

♻ ☆ SCA: Highly Efficient Semantic-Consistent Unrestricted Adversarial Attack

Zihao Pan, Weibin Wu, Yuhang Cao, Zibin Zheng

Unrestricted adversarial attacks typically manipulate the semantic content of an image (e.g., color or texture) to create adversarial examples that are both effective and photorealistic. Recent works have utilized the diffusion inversion process to map images into a latent space, where high-level semantics are manipulated by introducing perturbations. However, they often results in substantial semantic distortions in the denoised output and suffers from low efficiency. In this study, we propose a novel framework called Semantic-Consistent Unrestricted Adversarial Attacks (SCA), which employs an inversion method to extract edit-friendly noise maps and utilizes Multimodal Large Language Model (MLLM) to provide semantic guidance throughout the process. Under the condition of rich semantic information provided by MLLM, we perform the DDPM denoising process of each step using a series of edit-friendly noise maps, and leverage DPM Solver++ to accelerate this process, enabling efficient sampling with semantic consistency. Compared to existing methods, our framework enables the efficient generation of adversarial examples that exhibit minimal discernible semantic changes. Consequently, we for the first time introduce Semantic-Consistent Adversarial Examples (SCAE). Extensive experiments and visualizations have demonstrated the high efficiency of SCA, particularly in being on average 12 times faster than the state-of-the-art attacks. Our code can be found at https://github.com/Pan-Zihao/SCA.

♻ ☆ From Pixels to Tokens: Byte-Pair Encoding on Quantized Visual Modalities

Wanpeng Zhang, Zilong Xie, Yicheng Feng, Yijiang Li, Xingrun Xing, Sipeng Zheng, Zongqing Lu

♻ ☆ Semi-Supervised Fine-Tuning of Vision Foundation Models with Content-Style Decomposition

Mariia Drozdova, Vitaliy Kinakh, Yury Belousov, Erica Lastufka, Slava Voloshynovskiy

In this paper, we present a semi-supervised fine-tuning approach designed to improve the performance of pre-trained foundation models on downstream tasks with limited labeled data. By leveraging content-style decomposition within an information-theoretic framework, our method enhances the latent representations of pre-trained vision foundation models, aligning them more effectively with specific task objectives and addressing the problem of distribution shift. We evaluate our approach on multiple datasets, including MNIST, its augmented variations (with yellow and white stripes), CIFAR-10, SVHN, and GalaxyMNIST. The experiments show improvements over supervised finetuning baseline of pre-trained models, particularly in low-labeled data regimes, across both frozen and trainable backbones for the majority of the tested datasets.

comment: preprint

♻ ☆ DiffusionPID: Interpreting Diffusion via Partial Information Decomposition

Shaurya Dewan, Rushikesh Zawar, Prakanshul Saxena, Yingshan Chang, Andrew Luo, Yonatan Bisk

Text-to-image diffusion models have made significant progress in generating naturalistic images from textual inputs, and demonstrate the capacity to learn and represent complex visual-semantic relationships. While these diffusion models have achieved remarkable success, the underlying mechanisms driving their performance are not yet fully accounted for, with many unanswered questions surrounding what they learn, how they represent visual-semantic relationships, and why they sometimes fail to generalize. Our work presents Diffusion Partial Information Decomposition (DiffusionPID), a novel technique that applies information-theoretic principles to decompose the input text prompt into its elementary components, enabling a detailed examination of how individual tokens and their interactions shape the generated image. We introduce a formal approach to analyze the uniqueness, redundancy, and synergy terms by applying PID to the denoising model at both the image and pixel level. This approach enables us to characterize how individual tokens and their interactions affect the model output. We first present a fine-grained analysis of characteristics utilized by the model to uniquely localize specific concepts, we then apply our approach in bias analysis and show it can recover gender and ethnicity biases. Finally, we use our method to visually characterize word ambiguity and similarity from the model's perspective and illustrate the efficacy of our method for prompt intervention. Our results show that PID is a potent tool for evaluating and diagnosing text-to-image diffusion models.

♻ ☆ Scattering Spectra Models for Physics

Sihao Cheng, Rudy Morel, Erwan Allys, Brice Ménard, Stéphane Mallat

Physicists routinely need probabilistic models for a number of tasks such as parameter inference or the generation of new realizations of a field. Establishing such models for highly non-Gaussian fields is a challenge, especially when the number of samples is limited. In this paper, we introduce scattering spectra models for stationary fields and we show that they provide accurate and robust statistical descriptions of a wide range of fields encountered in physics. These models are based on covariances of scattering coefficients, i.e. wavelet decomposition of a field coupled with a point-wise modulus. After introducing useful dimension reductions taking advantage of the regularity of a field under rotation and scaling, we validate these models on various multi-scale physical fields and demonstrate that they reproduce standard statistics, including spatial moments up to 4th order. These scattering spectra provide us with a low-dimensional structured representation that captures key properties encountered in a wide range of physical fields. These generic models can be used for data exploration, classification, parameter inference, symmetry detection, and component separation.

comment: 11 pages, 6 figures, plus appendices, updated to published version

♻ ☆ AID: Attention Interpolation of Text-to-Image Diffusion NeurIPS 2024

Qiyuan He, Jinghao Wang, Ziwei Liu, Angela Yao

Conditional diffusion models can create unseen images in various settings, aiding image interpolation. Interpolation in latent spaces is well-studied, but interpolation with specific conditions like text or poses is less understood. Simple approaches, such as linear interpolation in the space of conditions, often result in images that lack consistency, smoothness, and fidelity. To that end, we introduce a novel training-free technique named Attention Interpolation via Diffusion (AID). Our key contributions include 1) proposing an inner/outer interpolated attention layer; 2) fusing the interpolated attention with self-attention to boost fidelity; and 3) applying beta distribution to selection to increase smoothness. We also present a variant, Prompt-guided Attention Interpolation via Diffusion (PAID), that considers interpolation as a condition-dependent generative process. This method enables the creation of new images with greater consistency, smoothness, and efficiency, and offers control over the exact path of interpolation. Our approach demonstrates effectiveness for conceptual and spatial interpolation. Code and demo are available at https://github.com/QY-H00/attention-interpolation-diffusion.

comment: NeurIPS 2024 Conference Paper

♻ ☆ Unraveling the Truth: Do VLMs really Understand Charts? A Deep Dive into Consistency and Robustness

Srija Mukhopadhyay, Adnan Qidwai, Aparna Garimella, Pritika Ramu, Vivek Gupta, Dan Roth

comment: 22 pages, 9 Tables, 5 figures, 22 examples

♻ ☆ Resfusion: Denoising Diffusion Probabilistic Models for Image Restoration Based on Prior Residual Noise NeurIPS 2024

Zhenning Shi, Haoshuai Zheng, Chen Xu, Changsheng Dong, Bin Pan, Xueshuo Xie, Along He, Tao Li, Huazhu Fu

Recently, research on denoising diffusion models has expanded its application to the field of image restoration. Traditional diffusion-based image restoration methods utilize degraded images as conditional input to effectively guide the reverse generation process, without modifying the original denoising diffusion process. However, since the degraded images already include low-frequency information, starting from Gaussian white noise will result in increased sampling steps. We propose Resfusion, a general framework that incorporates the residual term into the diffusion forward process, starting the reverse process directly from the noisy degraded images. The form of our inference process is consistent with the DDPM. We introduced a weighted residual noise, named resnoise, as the prediction target and explicitly provide the quantitative relationship between the residual term and the noise term in resnoise. By leveraging a smooth equivalence transformation, Resfusion determine the optimal acceleration step and maintains the integrity of existing noise schedules, unifying the training and inference processes. The experimental results demonstrate that Resfusion exhibits competitive performance on ISTD dataset, LOL dataset and Raindrop dataset with only five sampling steps. Furthermore, Resfusion can be easily applied to image generation and emerges with strong versatility. Our code and model are available at https://github.com/nkicsl/Resfusion.

comment: NeurIPS 2024

♻ ☆ Data Diet: Can Trimming PET/CT Datasets Enhance Lesion Segmentation?

Alexander Jaus, Simon Reiß, Jens Klesiek, Rainer Stiefelhagen

In this work, we describe our approach to compete in the autoPET3 datacentric track. While conventional wisdom suggests that larger datasets lead to better model performance, recent studies indicate that excluding certain training samples can enhance model accuracy. We find that in the autoPETIII dataset, a model that is trained on the entire dataset exhibits undesirable characteristics by producing a large number of false positives particularly for PSMA-PETs. We counteract this by removing the easiest samples from the training dataset as measured by the model loss before retraining from scratch. Using the proposed approach we manage to drive down the false negative volume and improve upon the baseline model in both false negative volume and dice score on the preliminary test set. Code and pre-trained models are available at github.com/alexanderjaus/autopet3_datadiet.

♻ ☆ Lessons Learned from a Unifying Empirical Study of Parameter-Efficient Transfer Learning (PETL) in Visual Recognition

Zheda Mai, Ping Zhang, Cheng-Hao Tu, Hong-You Chen, Li Zhang, Wei-Lun Chao

Parameter-efficient transfer learning (PETL) has attracted significant attention lately, due to the increasing size of pre-trained models and the need to fine-tune (FT) them for superior downstream performance. This community-wide enthusiasm has sparked a plethora of approaches. Nevertheless, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like when to apply PETL and which approach to use largely unanswered. In this paper, we conduct a unifying empirical study of representative PETL methods in the context of Vision Transformers. We systematically tune their hyper-parameters to fairly compare their accuracy on downstream tasks. Our study not only offers a valuable user guide but also unveils several new insights. First, if tuned carefully, different PETL methods can obtain similar accuracy in the low-shot benchmark VTAB-1K. This includes simple methods like FT the bias terms that were reported inferior. Second, though with similar accuracy, we find that PETL methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementariness) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PETL is also useful in many-shot regimes -- it achieves comparable and sometimes better accuracy than full FT, using much fewer learnable parameters. Last but not least, we investigate PETL's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., a CLIP backbone). Perhaps not surprisingly, PETL methods outperform full FT alone. However, with weight-space ensembles, the fully fine-tuned model can better balance target (i.e., downstream) distribution and distribution shift performance, suggesting a future research direction for PETL.

comment: Code is available at https://github.com/OSU-MLB/PETL_Vision

♻ ☆ VideoCLIP-XL: Advancing Long Description Understanding for Video CLIP Models EMNLP 2024

Jiapeng Wang, Chengyu Wang, Kunzhe Huang, Jun Huang, Lianwen Jin

comment: EMNLP 2024 Main conference

♻ ☆ CSIM: A Copula-based similarity index sensitive to local changes for Image quality assessment

Safouane El Ghazouali, Umberto Michelucci, Yassin El Hillali, Hichem Nouira

Image similarity metrics play an important role in computer vision applications, as they are used in image processing, computer vision and machine learning. Furthermore, those metrics enable tasks such as image retrieval, object recognition and quality assessment, essential in fields like healthcare, astronomy and surveillance. Existing metrics, such as PSNR, MSE, SSIM, ISSM and FSIM, often face limitations in terms of either speed, complexity or sensitivity to small changes in images. To address these challenges, a novel image similarity metric, namely CSIM, that combines real-time while being sensitive to subtle image variations is investigated in this paper. The novel metric uses Gaussian Copula from probability theory to transform an image into vectors of pixel distribution associated to local image patches. These vectors contain, in addition to intensities and pixel positions, information on the dependencies between pixel values, capturing the structural relationships within the image. By leveraging the properties of Copulas, CSIM effectively models the joint distribution of pixel intensities, enabling a more nuanced comparison of image patches making it more sensitive to local changes compared to other metrics. Experimental results demonstrate that CSIM outperforms existing similarity metrics in various image distortion scenarios, including noise, compression artifacts and blur. The metric's ability to detect subtle differences makes it suitable for applications requiring high precision, such as medical imaging, where the detection of minor anomalies can be of a high importance. The results obtained in this work can be reproduced from this Github repository: https://github.com/safouaneelg/copulasimilarity.

comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

♻ ☆ Grounding Continuous Representations in Geometry: Equivariant Neural Fields

David R Wessels, David M Knigge, Samuele Papa, Riccardo Valperga, Sharvaree Vadgama, Efstratios Gavves, Erik J Bekkers

Conditional Neural Fields (CNFs) are increasingly being leveraged as continuous signal representations, by associating each data-sample with a latent variable that conditions a shared backbone Neural Field (NeF) to reconstruct the sample. However, existing CNF architectures face limitations when using this latent downstream in tasks requiring fine grained geometric reasoning, such as classification and segmentation. We posit that this results from lack of explicit modelling of geometric information (e.g. locality in the signal or the orientation of a feature) in the latent space of CNFs. As such, we propose Equivariant Neural Fields (ENFs), a novel CNF architecture which uses a geometry-informed cross-attention to condition the NeF on a geometric variable, a latent point cloud of features, that enables an equivariant decoding from latent to field. We show that this approach induces a steerability property by which both field and latent are grounded in geometry and amenable to transformation laws: if the field transforms, the latent representation transforms accordingly - and vice versa. Crucially, this equivariance relation ensures that the latent is capable of (1) representing geometric patterns faitfhully, allowing for geometric reasoning in latent space, (2) weight-sharing over similar local patterns, allowing for efficient learning of datasets of fields. We validate these main properties in a range of tasks including classification, segmentation, forecasting and reconstruction, showing clear improvement over baselines with a geometry-free latent space.

♻ ☆ DiffIR2VR-Zero: Zero-Shot Video Restoration with Diffusion-based Image Restoration Models

Chang-Han Yeh, Chin-Yang Lin, Zhixiang Wang, Chi-Wei Hsiao, Ting-Hsuan Chen, Hau-Shiang Shiu, Yu-Lun Liu

This paper introduces a method for zero-shot video restoration using pre-trained image restoration diffusion models. Traditional video restoration methods often need retraining for different settings and struggle with limited generalization across various degradation types and datasets. Our approach uses a hierarchical token merging strategy for keyframes and local frames, combined with a hybrid correspondence mechanism that blends optical flow and feature-based nearest neighbor matching (latent merging). We show that our method not only achieves top performance in zero-shot video restoration but also significantly surpasses trained models in generalization across diverse datasets and extreme degradations (8$\times$ super-resolution and high-standard deviation video denoising). We present evidence through quantitative metrics and visual comparisons on various challenging datasets. Additionally, our technique works with any 2D restoration diffusion model, offering a versatile and powerful tool for video enhancement tasks without extensive retraining. This research leads to more efficient and widely applicable video restoration technologies, supporting advancements in fields that require high-quality video output. See our project page for video results and source code at https://jimmycv07.github.io/DiffIR2VR_web/.

comment: Project page: https://jimmycv07.github.io/DiffIR2VR_web/

♻ ☆ DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation

Chun-Hung Wu, Shih-Hong Chen, Chih-Yao Hu, Hsin-Yu Wu, Kai-Hsin Chen, Yu-You Chen, Chih-Hai Su, Chih-Kuo Lee, Yu-Lun Liu

This paper presents Deformable Neural Vessel Representations (DeNVeR), an unsupervised approach for vessel segmentation in X-ray videos without annotated ground truth. DeNVeR uses optical flow and layer separation, enhancing segmentation accuracy and adaptability through test-time training. A key component of our research is the introduction of the XACV dataset, the first X-ray angiography coronary video dataset with high-quality, manually labeled segmentation ground truth. Our evaluation demonstrates that DeNVeR outperforms current state-of-the-art methods in vessel segmentation. This paper marks an advance in medical imaging, providing a robust, data-efficient tool for disease diagnosis and treatment planning and setting a new standard for future research in video vessel segmentation. See our project page for video results at https://kirito878.github.io/DeNVeR/.

comment: Project page: https://kirito878.github.io/DeNVeR/

♻ ☆ A large-scale operational study of fingerprint quality and demographics

Javier Galbally, Aleksandrs Cepilovs, Ramon Blanco-Gonzalo, Gillian Ormiston, Oscar Miguel-Hurtado, Istvan Sz. Racz

Even though a few initial works have shown on small sets of data some level of bias in the performance of fingerprint recognition technology with respect to certain demographic groups, there is still not sufficient evidence to understand the impact that certain factors such as gender, age or finger-type may have on fingerprint quality and, in turn, also on fingerprint matching accuracy. The present work addresses this still under researched topic, on a large-scale database of operational data containing 10-print impressions of almost 16,000 subjects. The results reached provide further insight into the dependency of fingerprint quality and demographics, and show that there in fact exists a certain degree of performance variability in fingerprint-based recognition systems for different segments of the population. Based on the experimental evaluation, the work points out new observations based on data-driven evidence, provides plausible hypotheses to explain such observations, and concludes with potential follow-up actions that can help to reduce the observed fingerprint quality differences. This way, the current paper can be considered as a contribution to further increase the algorithmic fairness and equality of biometric technology.

comment: Extended journal version submitted to IET Biometrics. 10 pages, 5 figures Reference conference paper: J. Galbally, A. Cepilovs, R. Blanco-Gonzalo, G. Ormiston, O. Miguel-Hurtado, and I. S. Racz, 'Fingerprint quality per individual finger type: A large-scale study on real operational data' in Proc. IEEE Intl. Workshop on Biometrics and Forensics 2023 (IWBF 2023)

♻ ☆ Editable Concept Bottleneck Models

Lijie Hu, Chenyang Ren, Zhengyu Hu, Hongbin Lin, Cheng-Long Wang, Hui Xiong, Jingfeng Zhang, Di Wang

Concept Bottleneck Models (CBMs) have garnered much attention for their ability to elucidate the prediction process through a human-understandable concept layer. However, most previous studies focused on cases where the data, including concepts, are clean. In many scenarios, we always need to remove/insert some training data or new concepts from trained CBMs due to different reasons, such as privacy concerns, data mislabelling, spurious concepts, and concept annotation errors. Thus, the challenge of deriving efficient editable CBMs without retraining from scratch persists, particularly in large-scale applications. To address these challenges, we propose Editable Concept Bottleneck Models (ECBMs). Specifically, ECBMs support three different levels of data removal: concept-label-level, concept-level, and data-level. ECBMs enjoy mathematically rigorous closed-form approximations derived from influence functions that obviate the need for re-training. Experimental results demonstrate the efficiency and effectiveness of our ECBMs, affirming their adaptability within the realm of CBMs.

comment: 36 pages

♻ ☆ Qihoo-T2X: An Efficient Proxy-Tokenized Diffusion Transformer for Text-to-Any-Task

Jing Wang, Ao Ma, Jiasong Feng, Dawei Leng, Yuhui Yin, Xiaodan Liang

The global self-attention mechanism in diffusion transformers involves redundant computation due to the sparse and redundant nature of visual information, and the attention map of tokens within a spatial window shows significant similarity. To address this redundancy, we propose the Proxy-Tokenized Diffusion Transformer (PT-DiT), which employs sparse representative token attention (where the number of representative tokens is much smaller than the total number of tokens) to model global visual information efficiently. Specifically, within each transformer block, we compute an averaging token from each spatial-temporal window to serve as a proxy token for that region. The global semantics are captured through the self-attention of these proxy tokens and then injected into all latent tokens via cross-attention. Simultaneously, we introduce window and shift window attention to address the limitations in detail modeling caused by the sparse attention mechanism. Building on the well-designed PT-DiT, we further develop the Qihoo-T2X family, which includes a variety of models for T2I, T2V, and T2MV tasks. Experimental results show that PT-DiT achieves competitive performance while reducing the computational complexity in both image and video generation tasks (e.g., a 49% reduction compared to DiT and a 34% reduction compared to PixArt-$\alpha$). The visual exhibition and source code of Qihoo-T2X is available at https://360cvgroup.github.io/Qihoo-T2X/.

♻ ☆ DiffSF: Diffusion Models for Scene Flow Estimation

Yushan Zhang, Bastian Wandt, Maria Magnusson, Michael Felsberg

Scene flow estimation is an essential ingredient for a variety of real-world applications, especially for autonomous agents, such as self-driving cars and robots. While recent scene flow estimation approaches achieve a reasonable accuracy, their applicability to real-world systems additionally benefits from a reliability measure. Aiming at improving accuracy while additionally providing an estimate for uncertainty, we propose DiffSF that combines transformer-based scene flow estimation with denoising diffusion models. In the diffusion process, the ground truth scene flow vector field is gradually perturbed by adding Gaussian noise. In the reverse process, starting from randomly sampled Gaussian noise, the scene flow vector field prediction is recovered by conditioning on a source and a target point cloud. We show that the diffusion process greatly increases the robustness of predictions compared to prior approaches resulting in state-of-the-art performance on standard scene flow estimation benchmarks. Moreover, by sampling multiple times with different initial states, the denoising process predicts multiple hypotheses, which enables measuring the output uncertainty, allowing our approach to detect a majority of the inaccurate predictions. The code is available at https://github.com/ZhangYushan3/DiffSF.

♻ ☆ Bayesian Unsupervised Disentanglement of Anatomy and Geometry for Deep Groupwise Image Registration

Xinzhe Luo, Xin Wang, Linda Shapiro, Chun Yuan, Jianfeng Feng, Xiahai Zhuang

This article presents a general Bayesian learning framework for multi-modal groupwise image registration. The method builds on probabilistic modelling of the image generative process, where the underlying common anatomy and geometric variations of the observed images are explicitly disentangled as latent variables. Therefore, groupwise image registration is achieved via hierarchical Bayesian inference. We propose a novel hierarchical variational auto-encoding architecture to realise the inference procedure of the latent variables, where the registration parameters can be explicitly estimated in a mathematically interpretable fashion. Remarkably, this new paradigm learns groupwise image registration in an unsupervised closed-loop self-reconstruction process, sparing the burden of designing complex image-based similarity measures. The computationally efficient disentangled network architecture is also inherently scalable and flexible, allowing for groupwise registration on large-scale image groups with variable sizes. Furthermore, the inferred structural representations from multi-modal images via disentanglement learning are capable of capturing the latent anatomy of the observations with visual semantics. Extensive experiments were conducted to validate the proposed framework, including four different datasets from cardiac, brain, and abdominal medical images. The results have demonstrated the superiority of our method over conventional similarity-based approaches in terms of accuracy, efficiency, scalability, and interpretability.

♻ ☆ T-FAKE: Synthesizing Thermal Images for Facial Landmarking

Philipp Flotho, Moritz Piening, Anna Kukleva, Gabriele Steidl

Facial analysis is a key component in a wide range of applications such as security, autonomous driving, entertainment, and healthcare. Despite the availability of various facial RGB datasets, the thermal modality, which plays a crucial role in life sciences, medicine, and biometrics, has been largely overlooked. To address this gap, we introduce the T-FAKE dataset, a new large-scale synthetic thermal dataset with sparse and dense landmarks. To facilitate the creation of the dataset, we propose a novel RGB2Thermal loss function, which enables the transfer of thermal style to RGB faces. By utilizing the Wasserstein distance between thermal and RGB patches and the statistical analysis of clinical temperature distributions on faces, we ensure that the generated thermal images closely resemble real samples. Using RGB2Thermal style transfer based on our RGB2Thermal loss function, we create the T-FAKE dataset, a large-scale synthetic thermal dataset of faces. Leveraging our novel T-FAKE dataset, probabilistic landmark prediction, and label adaptation networks, we demonstrate significant improvements in landmark detection methods on thermal images across different landmark conventions. Our models show excellent performance with both sparse 70-point landmarks and dense 478-point landmark annotations. Our code and models are available at https://github.com/phflot/tfake.

comment: 22 pages, 12 figures, Philipp Flotho and Moritz Piening share equal contribution

♻ ☆ RobMOT: Robust 3D Multi-Object Tracking by Observational Noise and State Estimation Drift Mitigation on LiDAR PointCloud

Mohamed Nagy, Naoufel Werghi, Bilal Hassan, Jorge Dias, Majid Khonji

This work addresses limitations in recent 3D tracking-by-detection methods, focusing on identifying legitimate trajectories and addressing state estimation drift in Kalman filters. Current methods rely heavily on threshold-based filtering of false positive detections using detection scores to prevent ghost trajectories. However, this approach is inadequate for distant and partially occluded objects, where detection scores tend to drop, potentially leading to false positives exceeding the threshold. Additionally, the literature generally treats detections as precise localizations of objects. Our research reveals that noise in detections impacts localization information, causing trajectory drift for occluded objects and hindering recovery. To this end, we propose a novel online track validity mechanism that temporally distinguishes between legitimate and ghost tracks, along with a multi-stage observational gating process for incoming observations. This mechanism significantly improves tracking performance, with a $6.28\%$ in HOTA and a $17.87\%$ increase in MOTA. We also introduce a refinement to the Kalman filter that enhances noise mitigation in trajectory drift, leading to more robust state estimation for occluded objects. Our framework, RobMOT, outperforms state-of-the-art methods, including deep learning approaches, across various detectors, achieving up to a $4\%$ margin in HOTA and $6\%$ in MOTA. RobMOT excels under challenging conditions, such as prolonged occlusions and tracking distant objects, with up to a 59\% improvement in processing latency.

♻ ☆ Deep Model Interpretation with Limited Data : A Coreset-based Approach

Hamed Behzadi-Khormouji, José Oramas

Model Interpretation aims at the extraction of insights from the internals of a trained model. A common approach to address this task is the characterization of relevant features internally encoded in the model that are critical for its proper operation. Despite recent progress of these methods, they come with the weakness of being computationally expensive due to the dense evaluation of datasets that they require. As a consequence, research on the design of these methods have focused on smaller data subsets which may led to reduced insights. To address these computational costs, we propose a coreset-based interpretation framework that utilizes coreset selection methods to sample a representative subset of the large dataset for the interpretation task. Towards this goal, we propose a similarity-based evaluation protocol to assess the robustness of model interpretation methods towards the amount data they take as input. Experiments considering several interpretation methods, DNN models, and coreset selection methods show the effectiveness of the proposed framework.

♻ ☆ Applying the Lower-Biased Teacher Model in Semi-Supervised Object Detection

Shuang Wang

I present the Lower Biased Teacher model, an enhancement of the Unbiased Teacher model, specifically tailored for semi-supervised object detection tasks. The primary innovation of this model is the integration of a localization loss into the teacher model, which significantly improves the accuracy of pseudo-label generation. By addressing key issues such as class imbalance and the precision of bounding boxes, the Lower Biased Teacher model demonstrates superior performance in object detection tasks. Extensive experiments on multiple semi-supervised object detection datasets show that the Lower Biased Teacher model not only reduces the pseudo-labeling bias caused by class imbalances but also mitigates errors arising from incorrect bounding boxes. As a result, the model achieves higher mAP scores and more reliable detection outcomes compared to existing methods. This research underscores the importance of accurate pseudo-label generation and provides a robust framework for future advancements in semi-supervised learning for object detection.

comment: 12pages,2 figures,2 tables, several fomulas. arXiv admin note: text overlap with arXiv:2102.09480 by other authors

♻ ☆ Infrared Small Target Detection in Satellite Videos: A New Dataset and A Novel Recurrent Feature Refinement Framework

Xinyi Ying, Li Liu, Zaipin Lin, Yangsi Shi, Yingqian Wang, Ruojing Li, Xu Cao, Boyang Li, Shilin Zhou

Multi-frame infrared small target (MIRST) detection in satellite videos is a long-standing, fundamental yet challenging task for decades, and the challenges can be summarized as: First, extremely small target size, highly complex clutters & noises, various satellite motions result in limited feature representation, high false alarms, and difficult motion analyses. Second, the lack of large-scale public available MIRST dataset in satellite videos greatly hinders the algorithm development. To address the aforementioned challenges, in this paper, we first build a large-scale dataset for MIRST detection in satellite videos (namely IRSatVideo-LEO), and then develop a recurrent feature refinement (RFR) framework as the baseline method. Specifically, IRSatVideo-LEO is a semi-simulated dataset with synthesized satellite motion, target appearance, trajectory and intensity, which can provide a standard toolbox for satellite video generation and a reliable evaluation platform to facilitate the algorithm development. For baseline method, RFR is proposed to be equipped with existing powerful CNN-based methods for long-term temporal dependency exploitation and integrated motion compensation & MIRST detection. Specifically, a pyramid deformable alignment (PDA) module and a temporal-spatial-frequency modulation (TSFM) module are proposed to achieve effective and efficient feature alignment, propagation, aggregation and refinement. Extensive experiments have been conducted to demonstrate the effectiveness and superiority of our scheme. The comparative results show that ResUNet equipped with RFR outperforms the state-of-the-art MIRST detection methods. Dataset and code are released at https://github.com/XinyiYing/RFR.

♻ ☆ Explainable Artifacts for Synthetic Western Blot Source Attribution

João Phillipe Cardenuto, Sara Mandelli, Daniel Moreira, Paolo Bestagini, Edward Delp, Anderson Rocha

Recent advancements in artificial intelligence have enabled generative models to produce synthetic scientific images that are indistinguishable from pristine ones, posing a challenge even for expert scientists habituated to working with such content. When exploited by organizations known as paper mills, which systematically generate fraudulent articles, these technologies can significantly contribute to the spread of misinformation about ungrounded science, potentially undermining trust in scientific research. While previous studies have explored black-box solutions, such as Convolutional Neural Networks, for identifying synthetic content, only some have addressed the challenge of generalizing across different models and providing insight into the artifacts in synthetic images that inform the detection process. This study aims to identify explainable artifacts generated by state-of-the-art generative models (e.g., Generative Adversarial Networks and Diffusion Models) and leverage them for open-set identification and source attribution (i.e., pointing to the model that created the image).

comment: Accepted in IEEE International Workshop on Information Forensics and Security - WIFS 2024, Rome, Italy

♻ ☆ Evidence-based Match-status-Aware Gait Recognition for Out-of-Gallery Gait Identification

Heming Du, Chen Liu, Ming Wang, Lincheng Li, Shunli Zhang, Xin Yu

Existing gait recognition methods typically identify individuals based on the similarity between probe and gallery samples. However, these methods often neglect the fact that the gallery may not contain identities corresponding to the probes, leading to incorrect recognition.To identify Out-of-Gallery (OOG) gait queries, we propose an Evidence-based Match-status-Aware Gait Recognition (EMA-GR) framework. Inspired by Evidential Deep Learning (EDL), EMA-GR is designed to quantify the uncertainty associated with the match status of recognition. Thus, EMA-GR identifies whether the probe has a counterpart in the gallery. Specifically, we adopt an evidence collector to gather match status evidence from a recognition result pair and parameterize a Dirichlet distribution over the gathered evidence, following the Dempster-Shafer Theory of Evidence (DST). We measure the uncertainty and predict the match status of the recognition results, and thus determine whether the probe is an OOG query.To the best of our knowledge, our method is the first attempt to tackle OOG queries in gait recognition. Moreover, EMA-GR is agnostic against gait recognition methods and improves the robustness against OOG queries. Extensive experiments demonstrate that our method achieves state-of-the-art performance on datasets with OOG queries, and can also generalize well to other identity-retrieval tasks. Importantly, our method surpasses existing state-of-the-art methods by a substantial margin, achieving a 51.26% improvement when the OOG query rate is around 50% on OUMVLP.

comment: We have withdrawn this manuscript from Arxiv following a consensus among co-authors to refine our approach and reevaluate the data presented. This decision is part of our commitment to ensuring the highest standards of accuracy and completeness in our published work

♻ ☆ SDCL: Students Discrepancy-Informed Correction Learning for Semi-supervised Medical Image Segmentation MICCAI 2024

Bentao Song, Qingfeng Wang

Semi-supervised medical image segmentation (SSMIS) has been demonstrated the potential to mitigate the issue of limited medical labeled data. However, confirmation and cognitive biases may affect the prevalent teacher-student based SSMIS methods due to erroneous pseudo-labels. To tackle this challenge, we improve the mean teacher approach and propose the Students Discrepancy-Informed Correction Learning (SDCL) framework that includes two students and one non-trainable teacher, which utilizes the segmentation difference between the two students to guide the self-correcting learning. The essence of SDCL is to identify the areas of segmentation discrepancy as the potential bias areas, and then encourage the model to review the correct cognition and rectify their own biases in these areas. To facilitate the bias correction learning with continuous review and rectification, two correction loss functions are employed to minimize the correct segmentation voxel distance and maximize the erroneous segmentation voxel entropy. We conducted experiments on three public medical image datasets: two 3D datasets (CT and MRI) and one 2D dataset (MRI). The results show that our SDCL surpasses the current State-of-the-Art (SOTA) methods by 2.57\%, 3.04\%, and 2.34\% in the Dice score on the Pancreas, LA, and ACDC datasets, respectively. In addition, the accuracy of our method is very close to the fully supervised method on the ACDC dataset, and even exceeds the fully supervised method on the Pancreas and LA dataset. (Code available at \url{https://github.com/pascalcpp/SDCL}).

comment: Accepted at MICCAI 2024

♻ ☆ HarmoniCa: Harmonizing Training and Inference for Better Feature Cache in Diffusion Transformer Acceleration

Yushi Huang, Zining Wang, Ruihao Gong, Jing Liu, Xinjie Zhang, Jun Zhang

Diffusion Transformers (DiTs) have gained prominence for outstanding scalability and extraordinary performance in generative tasks. However, their considerable inference costs impede practical deployment. The feature cache mechanism, which involves storing and retrieving redundant computations across timesteps, holds promise for reducing per-step inference time in diffusion models. Most existing caching methods for DiT are manually designed. Although the learning-based approach attempts to optimize strategies adaptively, it suffers from discrepancies between training and inference, which hampers both the performance and acceleration ratio. Upon detailed analysis, we pinpoint that these discrepancies primarily stem from two aspects: (1) Prior Timestep Disregard, where training ignores the effect of cache usage at earlier timesteps, and (2) Objective Mismatch, where the training target (align predicted noise in each timestep) deviates from the goal of inference (generate the high-quality image). To alleviate these discrepancies, we propose HarmoniCa, a novel method that Harmonizes training and inference with a novel learning-based Caching framework built upon Step-Wise Denoising Training (SDT) and Image Error Proxy-Guided Objective (IEPO). Compared to the traditional training paradigm, the newly proposed SDT maintains the continuity of the denoising process, enabling the model to leverage information from prior timesteps during training, similar to the way it operates during inference. Furthermore, we design IEPO, which integrates an efficient proxy mechanism to approximate the final image error caused by reusing the cached feature. Therefore, IEPO helps balance final image quality and cache utilization, resolving the issue of training that only considers the impact of cache usage on the predicted output at each timestep.

comment: Code will be released soon

♻ ☆ Simple Drop-in LoRA Conditioning on Attention Layers Will Improve Your Diffusion Model

Joo Young Choi, Jaesung R. Park, Inkyu Park, Jaewoong Cho, Albert No, Ernest K. Ryu

Current state-of-the-art diffusion models employ U-Net architectures containing convolutional and (qkv) self-attention layers. The U-Net processes images while being conditioned on the time embedding input for each sampling step and the class or caption embedding input corresponding to the desired conditional generation. Such conditioning involves scale-and-shift operations to the convolutional layers but does not directly affect the attention layers. While these standard architectural choices are certainly effective, not conditioning the attention layers feels arbitrary and potentially suboptimal. In this work, we show that simply adding LoRA conditioning to the attention layers without changing or tuning the other parts of the U-Net architecture improves the image generation quality. For example, a drop-in addition of LoRA conditioning to EDM diffusion model yields FID scores of 1.91/1.75 for unconditional and class-conditional CIFAR-10 generation, improving upon the baseline of 1.97/1.79.

♻ ☆ TGIF: Text-Guided Inpainting Forgery Dataset

Hannes Mareen, Dimitrios Karageorgiou, Glenn Van Wallendael, Peter Lambert, Symeon Papadopoulos

Digital image manipulation has become increasingly accessible and realistic with the advent of generative AI technologies. Recent developments allow for text-guided inpainting, making sophisticated image edits possible with minimal effort. This poses new challenges for digital media forensics. For example, diffusion model-based approaches could either splice the inpainted region into the original image, or regenerate the entire image. In the latter case, traditional image forgery localization (IFL) methods typically fail. This paper introduces the Text-Guided Inpainting Forgery (TGIF) dataset, a comprehensive collection of images designed to support the training and evaluation of image forgery localization and synthetic image detection (SID) methods. The TGIF dataset includes approximately 75k forged images, originating from popular open-source and commercial methods, namely SD2, SDXL, and Adobe Firefly. We benchmark several state-of-the-art IFL and SID methods on TGIF. Whereas traditional IFL methods can detect spliced images, they fail to detect regenerated inpainted images. Moreover, traditional SID may detect the regenerated inpainted images to be fake, but cannot localize the inpainted area. Finally, both IFL and SID methods fail when exposed to stronger compression, while they are less robust to modern compression algorithms, such as WEBP. In conclusion, this work demonstrates the inefficiency of state-of-the-art detectors on local manipulations performed by modern generative approaches, and aspires to help with the development of more capable IFL and SID methods. The dataset and code can be downloaded at https://github.com/IDLabMedia/tgif-dataset.

comment: 6 pages, accepted at IEEE WIFS 2024

♻ ☆ Mapping Degeneration Meets Label Evolution: Learning Infrared Small Target Detection with Single Point Supervision

Xinyi Ying, Li Liu, Yingqian Wang, Ruojing Li, Nuo Chen, Zaiping Lin, Weidong Sheng, Shilin Zhou

Training a convolutional neural network (CNN) to detect infrared small targets in a fully supervised manner has gained remarkable research interests in recent years, but is highly labor expensive since a large number of per-pixel annotations are required. To handle this problem, in this paper, we make the first attempt to achieve infrared small target detection with point-level supervision. Interestingly, during the training phase supervised by point labels, we discover that CNNs first learn to segment a cluster of pixels near the targets, and then gradually converge to predict groundtruth point labels. Motivated by this "mapping degeneration" phenomenon, we propose a label evolution framework named label evolution with single point supervision (LESPS) to progressively expand the point label by leveraging the intermediate predictions of CNNs. In this way, the network predictions can finally approximate the updated pseudo labels, and a pixel-level target mask can be obtained to train CNNs in an end-to-end manner. We conduct extensive experiments with insightful visualizations to validate the effectiveness of our method. Experimental results show that CNNs equipped with LESPS can well recover the target masks from corresponding point labels, {and can achieve over 70% and 95% of their fully supervised performance in terms of pixel-level intersection over union (IoU) and object-level probability of detection (Pd), respectively. Code is available at https://github.com/XinyiYing/LESPS.

♻ ★ VCR: Visual Caption Restoration

Tianyu Zhang, Suyuchen Wang, Lu Li, Ge Zhang, Perouz Taslakian, Sai Rajeswar, Jie Fu, Bang Liu, Yoshua Bengio

We introduce Visual Caption Restoration (VCR), a novel vision-language task that challenges models to accurately restore partially obscured texts using pixel-level hints within images. This task stems from the observation that text embedded in images is intrinsically different from common visual elements and natural language due to the need to align the modalities of vision, text, and text embedded in images. While numerous works have integrated text embedded in images into visual question-answering tasks, approaches to these tasks generally rely on optical character recognition or masked language modeling, thus reducing the task to mainly text-based processing. However, text-based processing becomes ineffective in VCR as accurate text restoration depends on the combined information from provided images, context, and subtle cues from the tiny exposed areas of masked texts. We develop a pipeline to generate synthetic images for the VCR task using image-caption pairs, with adjustable caption visibility to control the task difficulty. With this pipeline, we construct a dataset for VCR called VCR-Wiki using images with captions from Wikipedia, comprising 2.11M English and 346K Chinese entities in both easy and hard split variants. Our results reveal that current vision language models significantly lag behind human performance in the VCR task, and merely fine-tuning the models on our dataset does not lead to notable improvements. We release VCR-Wiki and the data construction code to facilitate future research.

comment: 22 pages, 6 figures, 7 tables

♻ ☆ MITS-GAN: Safeguarding Medical Imaging from Tampering with Generative Adversarial Networks

Giovanni Pasqualino, Luca Guarnera, Alessandro Ortis, Sebastiano Battiato

The progress in generative models, particularly Generative Adversarial Networks (GANs), opened new possibilities for image generation but raised concerns about potential malicious uses, especially in sensitive areas like medical imaging. This study introduces MITS-GAN, a novel approach to prevent tampering in medical images, with a specific focus on CT scans. The approach disrupts the output of the attacker's CT-GAN architecture by introducing finely tuned perturbations that are imperceptible to the human eye. Specifically, the proposed approach involves the introduction of appropriate Gaussian noise to the input as a protective measure against various attacks. Our method aims to enhance tamper resistance, comparing favorably to existing techniques. Experimental results on a CT scan demonstrate MITS-GAN's superior performance, emphasizing its ability to generate tamper-resistant images with negligible artifacts. As image tampering in medical domains poses life-threatening risks, our proactive approach contributes to the responsible and ethical use of generative models. This work provides a foundation for future research in countering cyber threats in medical imaging. Models and codes are publicly available on https://iplab.dmi.unict.it/MITS-GAN-2024/.

♻ ☆ A Lightweight Modular Framework for Low-Cost Open-Vocabulary Object Detection Training

Bilal Faye, Binta Sow, Hanane Azzag, Mustapha Lebbah

Object detection is a fundamental challenge in computer vision, centered on recognizing objects within images, with diverse applications in areas like image analysis, robotics, and autonomous vehicles. Although existing methods have achieved great success, they are often constrained by a fixed vocabulary of objects. To overcome this limitation, approaches like MDETR have redefined object detection by incorporating region-level vision-language pre-training, enabling open-vocabulary object detectors. However, these methods are computationally heavy due to the simultaneous training of large models for both vision and language representations. To address this, we introduce a lightweight framework that significantly reduces the number of parameters while preserving, or even improving, performance. Our solution is applied to MDETR, resulting in the development of Lightweight MDETR (LightMDETR), an optimized version of MDETR designed to enhance computational efficiency without sacrificing accuracy. The core of our approach involves freezing the MDETR backbone and training only the Universal Projection module (UP), which bridges vision and language representations. A learnable modality token parameter allows the UP to seamlessly switch between modalities. Evaluations on tasks like phrase grounding, referring expression comprehension, and segmentation show that LightMDETR not only reduces computational costs but also outperforms several state-of-the-art methods in terms of accuracy.

♻ ☆ Deepfake Detection: A Comprehensive Survey from the Reliability Perspective

Tianyi Wang, Xin Liao, Kam Pui Chow, Xiaodong Lin, Yinglong Wang

The mushroomed Deepfake synthetic materials circulated on the internet have raised a profound social impact on politicians, celebrities, and individuals worldwide. In this survey, we provide a thorough review of the existing Deepfake detection studies from the reliability perspective. We identify three reliability-oriented research challenges in the current Deepfake detection domain: transferability, interpretability, and robustness. Moreover, while solutions have been frequently addressed regarding the three challenges, the general reliability of a detection model has been barely considered, leading to the lack of reliable evidence in real-life usages and even for prosecutions on Deepfake-related cases in court. We, therefore, introduce a model reliability study metric using statistical random sampling knowledge and the publicly available benchmark datasets to review the reliability of the existing detection models on arbitrary Deepfake candidate suspects. Case studies are further executed to justify the real-life Deepfake cases including different groups of victims with the help of the reliably qualified detection models as reviewed in this survey. Reviews and experiments on the existing approaches provide informative discussions and future research directions for Deepfake detection.

comment: Accepted to ACM Computing Surveys

♻ ☆ MA-FSAR: Multimodal Adaptation of CLIP for Few-Shot Action Recognition

Jiazheng Xing, Chao Xu, Mengmeng Wang, Guang Dai, Baigui Sun, Yong Liu, Jingdong Wang, Jian Zhao

Applying large-scale vision-language pre-trained models like CLIP to few-shot action recognition (FSAR) can significantly enhance both performance and efficiency. While several studies have recognized this advantage, most of them resort to full-parameter fine-tuning to make CLIP's visual encoder adapt to the FSAR data, which not only costs high computations but also overlooks the potential of the visual encoder to engage in temporal modeling and focus on targeted semantics directly. To tackle these issues, we introduce MA-FSAR, a framework that employs the Parameter-Efficient Fine-Tuning (PEFT) technique to enhance the CLIP visual encoder in terms of action-related temporal and semantic representations. Our solution involves a Fine-grained Multimodal Adaptation, which is different from the previous attempts of PEFT in regular action recognition. Specifically, we first insert a Global Temporal Adaptation that only receives the class token to capture global motion cues efficiently. Then these outputs integrate with visual tokens to enhance local temporal dynamics by a Local Multimodal Adaptation, which incorporates text features unique to the FSAR support set branch to highlight fine-grained semantics related to actions. In addition to these token-level designs, we propose a prototype-level text-guided construction module to further enrich the temporal and semantic characteristics of video prototypes. Extensive experiments demonstrate our superior performance in various tasks using minor trainable parameters.

♻ ☆ Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration

Kang Liao, Zongsheng Yue, Zhouxia Wang, Chen Change Loy

Although learning-based image restoration methods have made significant progress, they still struggle with limited generalization to real-world scenarios due to the substantial domain gap caused by training on synthetic data. Existing methods address this issue by improving data synthesis pipelines, estimating degradation kernels, employing deep internal learning, and performing domain adaptation and regularization. Previous domain adaptation methods have sought to bridge the domain gap by learning domain-invariant knowledge in either feature or pixel space. However, these techniques often struggle to extend to low-level vision tasks within a stable and compact framework. In this paper, we show that it is possible to perform domain adaptation via the noise space using diffusion models. In particular, by leveraging the unique property of how auxiliary conditional inputs influence the multi-step denoising process, we derive a meaningful diffusion loss that guides the restoration model in progressively aligning both restored synthetic and real-world outputs with a target clean distribution. We refer to this method as denoising as adaptation. To prevent shortcuts during joint training, we present crucial strategies such as channel-shuffling layer and residual-swapping contrastive learning in the diffusion model. They implicitly blur the boundaries between conditioned synthetic and real data and prevent the reliance of the model on easily distinguishable features. Experimental results on three classical image restoration tasks, namely denoising, deblurring, and deraining, demonstrate the effectiveness of the proposed method.

comment: Project Page: https://kangliao929.github.io/projects/noise-da/

♻ ☆ VDebugger: Harnessing Execution Feedback for Debugging Visual Programs EMNLP 2024

Xueqing Wu, Zongyu Lin, Songyan Zhao, Te-Lin Wu, Pan Lu, Nanyun Peng, Kai-Wei Chang

Visual programs are executable code generated by large language models to address visual reasoning problems. They decompose complex questions into multiple reasoning steps and invoke specialized models for each step to solve the problems. However, these programs are prone to logic errors, with our preliminary evaluation showing that 58% of the total errors are caused by program logic errors. Debugging complex visual programs remains a major bottleneck for visual reasoning. To address this, we introduce VDebugger, a novel critic-refiner framework trained to localize and debug visual programs by tracking execution step by step. VDebugger identifies and corrects program errors leveraging detailed execution feedback, improving interpretability and accuracy. The training data is generated through an automated pipeline that injects errors into correct visual programs using a novel mask-best decoding technique. Evaluations on six datasets demonstrate VDebugger's effectiveness, showing performance improvements of up to 3.2% in downstream task accuracy. Further studies show VDebugger's ability to generalize to unseen tasks, bringing a notable improvement of 2.3% on the unseen COVR task. Code, data and models are made publicly available at https://github.com/shirley-wu/vdebugger/

comment: EMNLP 2024 Findings

♻ ☆ M2Distill: Multi-Modal Distillation for Lifelong Imitation Learning ICRA2025

Kaushik Roy, Akila Dissanayake, Brendan Tidd, Peyman Moghadam

Lifelong imitation learning for manipulation tasks poses significant challenges due to distribution shifts that occur in incremental learning steps. Existing methods often focus on unsupervised skill discovery to construct an ever-growing skill library or distillation from multiple policies, which can lead to scalability issues as diverse manipulation tasks are continually introduced and may fail to ensure a consistent latent space throughout the learning process, leading to catastrophic forgetting of previously learned skills. In this paper, we introduce M2Distill, a multi-modal distillation-based method for lifelong imitation learning focusing on preserving consistent latent space across vision, language, and action distributions throughout the learning process. By regulating the shifts in latent representations across different modalities from previous to current steps, and reducing discrepancies in Gaussian Mixture Model (GMM) policies between consecutive learning steps, we ensure that the learned policy retains its ability to perform previously learned tasks while seamlessly integrating new skills. Extensive evaluations on the LIBERO lifelong imitation learning benchmark suites, including LIBERO-OBJECT, LIBERO-GOAL, and LIBERO-SPATIAL, demonstrate that our method consistently outperforms prior state-of-the-art methods across all evaluated metrics.

comment: Submitted to ICRA2025

♻ ☆ AVG-LLaVA: A Large Multimodal Model with Adaptive Visual Granularity

Zhibin Lan, Liqiang Niu, Fandong Meng, Wenbo Li, Jie Zhou, Jinsong Su

Recently, when dealing with high-resolution images, dominant LMMs usually divide them into multiple local images and one global image, which will lead to a large number of visual tokens. In this work, we introduce AVG-LLaVA, an LMM that can adaptively select the appropriate visual granularity based on the input image and instruction. This approach not only reduces the number of visual tokens and speeds up inference, but also improves the overall model performance. Specifically, we introduce the following modules based on LLaVA-NeXT: (a) a visual granularity scaler that includes multiple pooling layers to obtain visual tokens with different granularities; (b) a visual granularity router, which includes a Transformer layer, an MLP layer, and a voter layer, used to select the appropriate visual granularity based on the image and instruction. Furthermore, we propose RGLF, a novel training paradigm that aims at aligning the granularity predicted by the router with the preferences of the LMM, without the need for additional manually annotated data. Extensive experiments and analysis show that AVG-LLaVA achieves superior performance across 11 benchmarks, as well as significantly reduces the number of visual tokens and speeds up inference (e.g., an 85.3% reduction in visual tokens and a 2.53$\times$ increase in inference speed on the AI2D benchmark).

comment: Preprint

♻ ☆ LLaRA: Supercharging Robot Learning Data for Vision-Language Policy

Xiang Li, Cristina Mata, Jongwoo Park, Kumara Kahatapitiya, Yoo Sung Jang, Jinghuan Shang, Kanchana Ranasinghe, Ryan Burgert, Mu Cai, Yong Jae Lee, Michael S. Ryoo

LLMs with visual inputs, i.e., Vision Language Models (VLMs), have the capacity to process state information as visual-textual prompts and respond with policy decisions in text. We propose LLaRA: Large Language and Robotics Assistant, a framework that formulates robot action policy as conversations and provides improved action outputs when trained with auxiliary data that complements policy learning. We first introduce an automated pipeline to generate conversation-style instruction tuning data from existing behavior cloning data. Then we enrich the dataset in a self-supervised fashion by formulating six auxiliary tasks. A VLM finetuned with the resulting collection of datasets can generate meaningful robot action policy decisions. Our experiments across multiple simulated and real-world environments demonstrate the state-of-the-art performance of the proposed LLaRA framework. The code, datasets, and pretrained models are available at https://github.com/LostXine/LLaRA.

♻ ☆ Multi-hypotheses Conditioned Point Cloud Diffusion for 3D Human Reconstruction from Occluded Images NeurIPS 2024

Donghwan Kim, Tae-Kyun Kim

3D human shape reconstruction under severe occlusion due to human-object or human-human interaction is a challenging problem. Parametric models i.e., SMPL(-X), which are based on the statistics across human shapes, can represent whole human body shapes but are limited to minimally-clothed human shapes. Implicit-function-based methods extract features from the parametric models to employ prior knowledge of human bodies and can capture geometric details such as clothing and hair. However, they often struggle to handle misaligned parametric models and inpaint occluded regions given a single RGB image. In this work, we propose a novel pipeline, MHCDIFF, Multi-hypotheses Conditioned Point Cloud Diffusion, composed of point cloud diffusion conditioned on probabilistic distributions for pixel-aligned detailed 3D human reconstruction under occlusion. Compared to previous implicit-function-based methods, the point cloud diffusion model can capture the global consistent features to generate the occluded regions, and the denoising process corrects the misaligned SMPL meshes. The core of MHCDIFF is extracting local features from multiple hypothesized SMPL(-X) meshes and aggregating the set of features to condition the diffusion model. In the experiments on CAPE and MultiHuman datasets, the proposed method outperforms various SOTA methods based on SMPL, implicit functions, point cloud diffusion, and their combined, under synthetic and real occlusions. Our code is publicly available at https://donghwankim0101.github.io/projects/mhcdiff/ .

comment: 17 pages, 7 figures, accepted NeurIPS 2024

♻ ☆ Losing Visual Needles in Image Haystacks: Vision Language Models are Easily Distracted in Short and Long Contexts EMNLP 2024

Aditya Sharma, Michael Saxon, William Yang Wang

We present LoCoVQA, a dynamic benchmark generator for evaluating long-context extractive reasoning in vision language models (VLMs). LoCoVQA augments test examples for mathematical reasoning, VQA, and character recognition tasks with increasingly long visual contexts composed of both in-distribution and out-of-distribution distractor images. Across these tasks, a diverse set of VLMs rapidly lose performance as the visual context length grows, often exhibiting a striking logarithmic decay trend. This test assesses how well VLMs can ignore irrelevant information when answering queries -- a task that is quite easy for language models (LMs) in the text domain -- demonstrating that current state-of-the-art VLMs lack this essential capability for many long-context applications.

comment: Findings of EMNLP 2024

♻ ☆ Data-centric Prediction Explanation via Kernelized Stein Discrepancy

Mahtab Sarvmaili, Hassan Sajjad, Ga Wu

Existing example-based prediction explanation methods often bridge test and training data points through the model's parameters or latent representations. While these methods offer clues to the causes of model predictions, they often exhibit innate shortcomings, such as incurring significant computational overhead or producing coarse-grained explanations. This paper presents a Highly-precise and Data-centric Explan}ation (HD-Explain) prediction explanation method that exploits properties of Kernelized Stein Discrepancy (KSD). Specifically, the KSD uniquely defines a parameterized kernel function for a trained model that encodes model-dependent data correlation. By leveraging the kernel function, one can identify training samples that provide the best predictive support to a test point efficiently. We conducted thorough analyses and experiments across multiple classification domains, where we show that HD-Explain outperforms existing methods from various aspects, including 1) preciseness (fine-grained explanation), 2) consistency, and 3) computation efficiency, leading to a surprisingly simple, effective, and robust prediction explanation solution.

♻ ☆ WcDT: World-centric Diffusion Transformer for Traffic Scene Generation

Chen Yang, Yangfan He, Aaron Xuxiang Tian, Dong Chen, Jianhui Wang, Tianyu Shi, Arsalan Heydarian

In this paper, we introduce a novel approach for autonomous driving trajectory generation by harnessing the complementary strengths of diffusion probabilistic models (a.k.a., diffusion models) and transformers. Our proposed framework, termed the "World-Centric Diffusion Transformer"(WcDT), optimizes the entire trajectory generation process, from feature extraction to model inference. To enhance the scene diversity and stochasticity, the historical trajectory data is first preprocessed into "Agent Move Statement" and encoded into latent space using Denoising Diffusion Probabilistic Models (DDPM) enhanced with Diffusion with Transformer (DiT) blocks. Then, the latent features, historical trajectories, HD map features, and historical traffic signal information are fused with various transformer-based encoders that are used to enhance the interaction of agents with other elements in the traffic scene. The encoded traffic scenes are then decoded by a trajectory decoder to generate multimodal future trajectories. Comprehensive experimental results show that the proposed approach exhibits superior performance in generating both realistic and diverse trajectories, showing its potential for integration into automatic driving simulation systems. Our code is available at \url{https://github.com/yangchen1997/WcDT}.

comment: 7 pages, 5 figures

♻ ☆ Single-Image Shadow Removal Using Deep Learning: A Comprehensive Survey

Laniqng Guo, Chong Wang, Yufei Wang, Yi Yu, Siyu Huang, Wenhan Yang, Alex C. Kot, Bihan Wen

Shadow removal aims at restoring the image content within shadow regions, pursuing a uniform distribution of illumination that is consistent between shadow and non-shadow regions. {Comparing to other image restoration tasks, there are two unique challenges in shadow removal:} 1) The patterns of shadows are arbitrary, varied, and often have highly complex trace structures, making ``trace-less'' image recovery difficult. 2) The degradation caused by shadows is spatially non-uniform, resulting in inconsistencies in illumination and color between shadow and non-shadow areas. Recent developments in this field are primarily driven by deep learning-based solutions, employing a variety of learning strategies, network architectures, loss functions, and training data. Nevertheless, a thorough and insightful review of deep learning-based shadow removal techniques is still lacking. In this paper, we are the first to provide a comprehensive survey to cover various aspects ranging from technical details to applications. We highlight the major advancements in deep learning-based single-image shadow removal methods, thoroughly review previous research across various categories, and provide insights into the historical progression of these developments. Additionally, we summarize performance comparisons both quantitatively and qualitatively. Beyond the technical aspects of shadow removal methods, we also explore potential future directions for this field.

comment: url: https://github.com/GuoLanqing/Awesome-Shadow-Removal

♻ ☆ Finer: Investigating and Enhancing Fine-Grained Visual Concept Recognition in Large Vision Language Models

Jeonghwan Kim, Heng Ji

Recent advances in instruction-tuned Large Vision-Language Models (LVLMs) have imbued the models with the ability to generate high-level, image-grounded explanations with ease. While such capability is largely attributed to the rich world knowledge contained within the Large Language Models (LLMs), our work reveals their shortcomings in fine-grained visual categorization (FGVC) across six different benchmark settings. Most recent state-of-the-art LVLMs like LLaVa-1.5, InstructBLIP and GPT-4V not only severely deteriorate in terms of classification performance, e.g., average drop of 65.58 in EM for Stanford Dogs for LLaVA-1.5, but also struggle to generate an accurate explanation with detailed attributes based on the concept that appears within an input image despite their capability to generate holistic image-level descriptions. In-depth analyses show that instruction-tuned LVLMs exhibit modality gap, showing discrepancy when given textual and visual inputs that correspond to the same concept, preventing the image modality from leveraging the rich parametric knowledge within the LLMs. In an effort to further the community's endeavor in this direction, we propose a multiple granularity attribute-centric evaluation benchmark, Finer, which aims to establish a ground to evaluate LVLMs' fine-grained visual comprehension ability and provide significantly improved explainability.

Machine Learning 152

☆ System 2 reasoning capabilities are nigh

Scott C. Lowe

In recent years, machine learning models have made strides towards human-like reasoning capabilities from several directions. In this work, we review the current state of the literature and describe the remaining steps to achieve a neural model which can perform System 2 reasoning analogous to a human. We argue that if current models are insufficient to be classed as performing reasoning, there remains very little additional progress needed to attain that goal.

☆ RAFT: Realistic Attacks to Fool Text Detectors EMNLP 2024

James Wang, Ran Li, Junfeng Yang, Chengzhi Mao

comment: Accepted by EMNLP 2024

☆ Geometric Representation Condition Improves Equivariant Molecule Generation

Zian Li, Cai Zhou, Xiyuan Wang, Xingang Peng, Muhan Zhang

Recent advancements in molecular generative models have demonstrated substantial potential in accelerating scientific discovery, particularly in drug design. However, these models often face challenges in generating high-quality molecules, especially in conditional scenarios where specific molecular properties must be satisfied. In this work, we introduce GeoRCG, a general framework to enhance the performance of molecular generative models by integrating geometric representation conditions. We decompose the molecule generation process into two stages: first, generating an informative geometric representation; second, generating a molecule conditioned on the representation. Compared to directly generating a molecule, the relatively easy-to-generate representation in the first-stage guides the second-stage generation to reach a high-quality molecule in a more goal-oriented and much faster way. Leveraging EDM as the base generator, we observe significant quality improvements in unconditional molecule generation on the widely-used QM9 and GEOM-DRUG datasets. More notably, in the challenging conditional molecular generation task, our framework achieves an average 31\% performance improvement over state-of-the-art approaches, highlighting the superiority of conditioning on semantically rich geometric representations over conditioning on individual property values as in previous approaches. Furthermore, we show that, with such representation guidance, the number of diffusion steps can be reduced to as small as 100 while maintaining superior generation quality than that achieved with 1,000 steps, thereby significantly accelerating the generation process.

☆ Learning Humanoid Locomotion over Challenging Terrain

Ilija Radosavovic, Sarthak Kamat, Trevor Darrell, Jitendra Malik

Humanoid robots can, in principle, use their legs to go almost anywhere. Developing controllers capable of traversing diverse terrains, however, remains a considerable challenge. Classical controllers are hard to generalize broadly while the learning-based methods have primarily focused on gentle terrains. Here, we present a learning-based approach for blind humanoid locomotion capable of traversing challenging natural and man-made terrain. Our method uses a transformer model to predict the next action based on the history of proprioceptive observations and actions. The model is first pre-trained on a dataset of flat-ground trajectories with sequence modeling, and then fine-tuned on uneven terrain using reinforcement learning. We evaluate our model on a real humanoid robot across a variety of terrains, including rough, deformable, and sloped surfaces. The model demonstrates robust performance, in-context adaptation, and emergent terrain representations. In real-world case studies, our humanoid robot successfully traversed over 4 miles of hiking trails in Berkeley and climbed some of the steepest streets in San Francisco.

comment: Project page: https://humanoid-challenging-terrain.github.io

☆ Minimax-optimal trust-aware multi-armed bandits

Changxiao Cai, Jiacheng Zhang

Multi-armed bandit (MAB) algorithms have achieved significant success in sequential decision-making applications, under the premise that humans perfectly implement the recommended policy. However, existing methods often overlook the crucial factor of human trust in learning algorithms. When trust is lacking, humans may deviate from the recommended policy, leading to undesired learning performance. Motivated by this gap, we study the trust-aware MAB problem by integrating a dynamic trust model into the standard MAB framework. Specifically, it assumes that the recommended and actually implemented policy differs depending on human trust, which in turn evolves with the quality of the recommended policy. We establish the minimax regret in the presence of the trust issue and demonstrate the suboptimality of vanilla MAB algorithms such as the upper confidence bound (UCB) algorithm. To overcome this limitation, we introduce a novel two-stage trust-aware procedure that provably attains near-optimal statistical guarantees. A simulation study is conducted to illustrate the benefits of our proposed algorithm when dealing with the trust issue.

☆ GenSim2: Scaling Robot Data Generation with Multi-modal and Reasoning LLMs

Pu Hua, Minghuan Liu, Annabella Macaluso, Yunfeng Lin, Weinan Zhang, Huazhe Xu, Lirui Wang

comment: CoRL 2024. Project website: https://gensim2.github.io/

☆ Real-World Benchmarks Make Membership Inference Attacks Fail on Diffusion Models

Chumeng Liang, Jiaxuan You

Membership inference attacks (MIAs) on diffusion models have emerged as potential evidence of unauthorized data usage in training pre-trained diffusion models. These attacks aim to detect the presence of specific images in training datasets of diffusion models. Our study delves into the evaluation of state-of-the-art MIAs on diffusion models and reveals critical flaws and overly optimistic performance estimates in existing MIA evaluation. We introduce CopyMark, a more realistic MIA benchmark that distinguishes itself through the support for pre-trained diffusion models, unbiased datasets, and fair evaluation pipelines. Through extensive experiments, we demonstrate that the effectiveness of current MIA methods significantly degrades under these more practical conditions. Based on our results, we alert that MIA, in its current state, is not a reliable approach for identifying unauthorized data usage in pre-trained diffusion models. To the best of our knowledge, we are the first to discover the performance overestimation of MIAs on diffusion models and present a unified benchmark for more realistic evaluation. Our code is available on GitHub: \url{https://github.com/caradryanl/CopyMark}.

☆ Conditional Enzyme Generation Using Protein Language Models with Adapters

Jason Yang, Aadyot Bhatnagar, Jeffrey A. Ruffolo, Ali Madani

The conditional generation of proteins with desired functions and/or properties is a key goal for generative models. Existing methods based on prompting of language models can generate proteins conditioned on a target functionality, such as a desired enzyme family. However, these methods are limited to simple, tokenized conditioning and have not been shown to generalize to unseen functions. In this study, we propose ProCALM (Protein Conditionally Adapted Language Model), an approach for the conditional generation of proteins using adapters to protein language models. Our specific implementation of ProCALM involves finetuning ProGen2 to incorporate conditioning representations of enzyme function and taxonomy. ProCALM matches existing methods at conditionally generating sequences from target enzyme families. Impressively, it can also generate within the joint distribution of enzymatic function and taxonomy, and it can generalize to rare and unseen enzyme families and taxonomies. Overall, ProCALM is a flexible and computationally efficient approach, and we expect that it can be extended to a wide range of generative language models.

☆ Robust Offline Imitation Learning from Diverse Auxiliary Data

Udita Ghosh, Dripta S. Raychaudhuri, Jiachen Li, Konstantinos Karydis, Amit K. Roy-Chowdhury

Offline imitation learning enables learning a policy solely from a set of expert demonstrations, without any environment interaction. To alleviate the issue of distribution shift arising due to the small amount of expert data, recent works incorporate large numbers of auxiliary demonstrations alongside the expert data. However, the performance of these approaches rely on assumptions about the quality and composition of the auxiliary data. However, they are rarely successful when those assumptions do not hold. To address this limitation, we propose Robust Offline Imitation from Diverse Auxiliary Data (ROIDA). ROIDA first identifies high-quality transitions from the entire auxiliary dataset using a learned reward function. These high-reward samples are combined with the expert demonstrations for weighted behavioral cloning. For lower-quality samples, ROIDA applies temporal difference learning to steer the policy towards high-reward states, improving long-term returns. This two-pronged approach enables our framework to effectively leverage both high and low-quality data without any assumptions. Extensive experiments validate that ROIDA achieves robust and consistent performance across multiple auxiliary datasets with diverse ratios of expert and non-expert demonstrations. ROIDA effectively leverages unlabeled auxiliary data, outperforming prior methods reliant on specific data assumptions.

☆ A Global Medical Data Security and Privacy Preserving Standards Identification Framework for Electronic Healthcare Consumers

Vinaytosh Mishra, Kishu Gupta, Deepika Saxena, Ashutosh Kumar Singh

Electronic Health Records (EHR) are crucial for the success of digital healthcare, with a focus on putting consumers at the center of this transformation. However, the digitalization of healthcare records brings along security and privacy risks for personal data. The major concern is that different countries have varying standards for the security and privacy of medical data. This paper proposed a novel and comprehensive framework to standardize these rules globally, bringing them together on a common platform. To support this proposal, the study reviews existing literature to understand the research interest in this issue. It also examines six key laws and standards related to security and privacy, identifying twenty concepts. The proposed framework utilized K-means clustering to categorize these concepts and identify five key factors. Finally, an Ordinal Priority Approach is applied to determine the preferred implementation of these factors in the context of EHRs. The proposed study provides a descriptive then prescriptive framework for the implementation of privacy and security in the context of electronic health records. Therefore, the findings of the proposed framework are useful for professionals and policymakers in improving the security and privacy associated with EHRs.

☆ Open-World Reinforcement Learning over Long Short-Term Imagination

Jiajian Li, Qi Wang, Yunbo Wang, Xin Jin, Yang Li, Wenjun Zeng, Xiaokang Yang

Training visual reinforcement learning agents in a high-dimensional open world presents significant challenges. While various model-based methods have improved sample efficiency by learning interactive world models, these agents tend to be "short-sighted", as they are typically trained on short snippets of imagined experiences. We argue that the primary obstacle in open-world decision-making is improving the efficiency of off-policy exploration across an extensive state space. In this paper, we present LS-Imagine, which extends the imagination horizon within a limited number of state transition steps, enabling the agent to explore behaviors that potentially lead to promising long-term feedback. The foundation of our approach is to build a long short-term world model. To achieve this, we simulate goal-conditioned jumpy state transitions and compute corresponding affordance maps by zooming in on specific areas within single images. This facilitates the integration of direct long-term values into behavior learning. Our method demonstrates significant improvements over state-of-the-art techniques in MineDojo.

☆ What Matters for Model Merging at Scale?

Prateek Yadav, Tu Vu, Jonathan Lai, Alexandra Chronopoulou, Manaal Faruqui, Mohit Bansal, Tsendsuren Munkhdalai

comment: 20 Pages, 7 Figures, 4 Tables

☆ Large Language Model Performance Benchmarking on Mobile Platforms: A Thorough Evaluation

Jie Xiao, Qianyi Huang, Xu Chen, Chen Tian

As large language models (LLMs) increasingly integrate into every aspect of our work and daily lives, there are growing concerns about user privacy, which push the trend toward local deployment of these models. There are a number of lightweight LLMs (e.g., Gemini Nano, LLAMA2 7B) that can run locally on smartphones, providing users with greater control over their personal data. As a rapidly emerging application, we are concerned about their performance on commercial-off-the-shelf mobile devices. To fully understand the current landscape of LLM deployment on mobile platforms, we conduct a comprehensive measurement study on mobile devices. We evaluate both metrics that affect user experience, including token throughput, latency, and battery consumption, as well as factors critical to developers, such as resource utilization, DVFS strategies, and inference engines. In addition, we provide a detailed analysis of how these hardware capabilities and system dynamics affect on-device LLM performance, which may help developers identify and address bottlenecks for mobile LLM applications. We also provide comprehensive comparisons across the mobile system-on-chips (SoCs) from major vendors, highlighting their performance differences in handling LLM workloads. We hope that this study can provide insights for both the development of on-device LLMs and the design for future mobile system architecture.

☆ TICKing All the Boxes: Generated Checklists Improve LLM Evaluation and Generation

Jonathan Cook, Tim Rocktäschel, Jakob Foerster, Dennis Aumiller, Alex Wang

☆ Exploring gauge-fixing conditions with gradient-based optimization

William Detmold, Gurtej Kanwar, Yin Lin, Phiala E. Shanahan, Michael L. Wagman

Lattice gauge fixing is required to compute gauge-variant quantities, for example those used in RI-MOM renormalization schemes or as objects of comparison for model calculations. Recently, gauge-variant quantities have also been found to be more amenable to signal-to-noise optimization using contour deformations. These applications motivate systematic parameterization and exploration of gauge-fixing schemes. This work introduces a differentiable parameterization of gauge fixing which is broad enough to cover Landau gauge, Coulomb gauge, and maximal tree gauges. The adjoint state method allows gradient-based optimization to select gauge-fixing schemes that minimize an arbitrary target loss function.

comment: 9 pages, 2 figures; Proceedings of the 41st International Symposium on Lattice Field Theory (Lattice 2024)

☆ How Discrete and Continuous Diffusion Meet: Comprehensive Analysis of Discrete Diffusion Models via a Stochastic Integral Framework

Yinuo Ren, Haoxuan Chen, Grant M. Rotskoff, Lexing Ying

Discrete diffusion models have gained increasing attention for their ability to model complex distributions with tractable sampling and inference. However, the error analysis for discrete diffusion models remains less well-understood. In this work, we propose a comprehensive framework for the error analysis of discrete diffusion models based on L\'evy-type stochastic integrals. By generalizing the Poisson random measure to that with a time-independent and state-dependent intensity, we rigorously establish a stochastic integral formulation of discrete diffusion models and provide the corresponding change of measure theorems that are intriguingly analogous to It\^o integrals and Girsanov's theorem for their continuous counterparts. Our framework unifies and strengthens the current theoretical results on discrete diffusion models and obtains the first error bound for the $\tau$-leaping scheme in KL divergence. With error sources clearly identified, our analysis gives new insight into the mathematical properties of discrete diffusion models and offers guidance for the design of efficient and accurate algorithms for real-world discrete diffusion model applications.

☆ Understanding Reasoning in Chain-of-Thought from the Hopfieldian View

Lijie Hu, Liang Liu, Shu Yang, Xin Chen, Zhen Tan, Muhammad Asif Ali, Mengdi Li, Di Wang

comment: 28 pages, a new version of "A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning"

☆ Training Over a Distribution of Hyperparameters for Enhanced Performance and Adaptability on Imbalanced Classification

Kelsey Lieberman, Swarna Kamlam Ravindran, Shuai Yuan, Carlo Tomasi

Although binary classification is a well-studied problem, training reliable classifiers under severe class imbalance remains a challenge. Recent techniques mitigate the ill effects of imbalance on training by modifying the loss functions or optimization methods. We observe that different hyperparameter values on these loss functions perform better at different recall values. We propose to exploit this fact by training one model over a distribution of hyperparameter values--instead of a single value--via Loss Conditional Training (LCT). Experiments show that training over a distribution of hyperparameters not only approximates the performance of several models but actually improves the overall performance of models on both CIFAR and real medical imaging applications, such as melanoma and diabetic retinopathy detection. Furthermore, training models with LCT is more efficient because some hyperparameter tuning can be conducted after training to meet individual needs without needing to retrain from scratch.

☆ Nonstationary Sparse Spectral Permanental Process

Zicheng Sun, Yixuan Zhang, Zenan Ling, Xuhui Fan, Feng Zhou

Existing permanental processes often impose constraints on kernel types or stationarity, limiting the model's expressiveness. To overcome these limitations, we propose a novel approach utilizing the sparse spectral representation of nonstationary kernels. This technique relaxes the constraints on kernel types and stationarity, allowing for more flexible modeling while reducing computational complexity to the linear level. Additionally, we introduce a deep kernel variant by hierarchically stacking multiple spectral feature mappings, further enhancing the model's expressiveness to capture complex patterns in data. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of our approach, particularly in scenarios with pronounced data nonstationarity. Additionally, ablation studies are conducted to provide insights into the impact of various hyperparameters on model performance.

☆ HyResPINNs: Adaptive Hybrid Residual Networks for Learning Optimal Combinations of Neural and RBF Components for Physics-Informed Modeling

Madison Cooley, Robert M. Kirby, Shandian Zhe, Varun Shankar

Physics-informed neural networks (PINNs) are an increasingly popular class of techniques for the numerical solution of partial differential equations (PDEs), where neural networks are trained using loss functions regularized by relevant PDE terms to enforce physical constraints. We present a new class of PINNs called HyResPINNs, which augment traditional PINNs with adaptive hybrid residual blocks that combine the outputs of a standard neural network and a radial basis function (RBF) network. A key feature of our method is the inclusion of adaptive combination parameters within each residual block, which dynamically learn to weigh the contributions of the neural network and RBF network outputs. Additionally, adaptive connections between residual blocks allow for flexible information flow throughout the network. We show that HyResPINNs are more robust to training point locations and neural network architectures than traditional PINNs. Moreover, HyResPINNs offer orders of magnitude greater accuracy than competing methods on certain problems, with only modest increases in training costs. We demonstrate the strengths of our approach on challenging PDEs, including the Allen-Cahn equation and the Darcy-Flow equation. Our results suggest that HyResPINNs effectively bridge the gap between traditional numerical methods and modern machine learning-based solvers.

comment: 14 pages, 6 figures

☆ Teaching Transformers Modular Arithmetic at Scale

Eshika Saxena, Alberto Alfarano, Emily Wenger, Kristin Lauter

Modular addition is, on its face, a simple operation: given $N$ elements in $\mathbb{Z}_q$, compute their sum modulo $q$. Yet, scalable machine learning solutions to this problem remain elusive: prior work trains ML models that sum $N \le 6$ elements mod $q \le 1000$. Promising applications of ML models for cryptanalysis-which often involve modular arithmetic with large $N$ and $q$-motivate reconsideration of this problem. This work proposes three changes to the modular addition model training pipeline: more diverse training data, an angular embedding, and a custom loss function. With these changes, we demonstrate success with our approach for $N = 256, q = 3329$, a case which is interesting for cryptographic applications, and a significant increase in $N$ and $q$ over prior work. These techniques also generalize to other modular arithmetic problems, motivating future work.

☆ Towards Linguistically-Aware and Language-Independent Tokenization for Large Language Models (LLMs)

Abrar Rahman, Garry Bowlin, Binit Mohanty, Sean McGunigal

☆ Training on more Reachable Tasks for Generalisation in Reinforcement Learning

Max Weltevrede, Caroline Horsch, Matthijs T. J. Spaan, Wendelin Böhmer

In multi-task reinforcement learning, agents train on a fixed set of tasks and have to generalise to new ones. Recent work has shown that increased exploration improves this generalisation, but it remains unclear why exactly that is. In this paper, we introduce the concept of reachability in multi-task reinforcement learning and show that an initial exploration phase increases the number of reachable tasks the agent is trained on. This, and not the increased exploration, is responsible for the improved generalisation, even to unreachable tasks. Inspired by this, we propose a novel method Explore-Go that implements such an exploration phase at the beginning of each episode. Explore-Go only modifies the way experience is collected and can be used with most existing on-policy or off-policy reinforcement learning algorithms. We demonstrate the effectiveness of our method when combined with some popular algorithms and show an increase in generalisation performance across several environments.

comment: arXiv admin note: text overlap with arXiv:2406.08069

☆ Ward: Provable RAG Dataset Inference via LLM Watermarks

Nikola Jovanović, Robin Staab, Maximilian Baader, Martin Vechev

Retrieval-Augmented Generation (RAG) improves LLMs by enabling them to incorporate external data during generation. This raises concerns for data owners regarding unauthorized use of their content in RAG systems. Despite its importance, the challenge of detecting such unauthorized usage remains underexplored, with existing datasets and methodologies from adjacent fields being ill-suited for its study. In this work, we take several steps to bridge this gap. First, we formalize this problem as (black-box) RAG Dataset Inference (RAG-DI). To facilitate research on this challenge, we further introduce a novel dataset specifically designed for benchmarking RAG-DI methods under realistic conditions, and propose a set of baseline approaches. Building on this foundation, we introduce Ward, a RAG-DI method based on LLM watermarks that enables data owners to obtain rigorous statistical guarantees regarding the usage of their dataset in a RAG system. In our experimental evaluation, we show that Ward consistently outperforms all baselines across many challenging settings, achieving higher accuracy, superior query efficiency and robustness. Our work provides a foundation for future studies of RAG-DI and highlights LLM watermarks as a promising approach to this problem.

☆ NRGBoost: Energy-Based Generative Boosted Trees

João Bravo

Despite the rise to dominance of deep learning in unstructured data domains, tree-based methods such as Random Forests (RF) and Gradient Boosted Decision Trees (GBDT) are still the workhorses for handling discriminative tasks on tabular data. We explore generative extensions of these popular algorithms with a focus on explicitly modeling the data density (up to a normalization constant), thus enabling other applications besides sampling. As our main contribution we propose an energy-based generative boosting algorithm that is analogous to the second order boosting implemented in popular packages like XGBoost. We show that, despite producing a generative model capable of handling inference tasks over any input variable, our proposed algorithm can achieve similar discriminative performance to GBDT on a number of real world tabular datasets, outperforming alternative generative approaches. At the same time, we show that it is also competitive with neural network based models for sampling.

☆ No Need to Talk: Asynchronous Mixture of Language Models

Anastasiia Filippova, Angelos Katharopoulos, David Grangier, Ronan Collobert

comment: 23 pages

☆ A Probabilistic Perspective on Unlearning and Alignment for Large Language Models

Yan Scholten, Stephan Günnemann, Leo Schwinn

Comprehensive evaluation of Large Language Models (LLMs) is an open research problem. Existing evaluations rely on deterministic point estimates generated via greedy decoding. However, we find that deterministic evaluations fail to capture the whole output distribution of a model, yielding inaccurate estimations of model capabilities. This is particularly problematic in critical contexts such as unlearning and alignment, where precise model evaluations are crucial. To remedy this, we introduce the first formal probabilistic evaluation framework in LLMs. Namely, we derive novel metrics with high-probability guarantees concerning the output distribution of a model. Our metrics are application-independent and allow practitioners to make more reliable estimates about model capabilities before deployment. Through a case study focused on unlearning, we reveal that deterministic evaluations falsely indicate successful unlearning, whereas our probabilistic evaluations demonstrate that most if not all of the supposedly unlearned information remains accessible in these models. Additionally, we propose a novel unlearning loss based on entropy optimization and adaptive temperature scaling, which significantly improves unlearning in probabilistic settings on recent benchmarks. Our proposed shift from point estimates to probabilistic evaluations of output distributions represents an important step toward comprehensive evaluations of LLMs. https://github.com/yascho/probabilistic-unlearning

☆ Improving Online Bagging for Complex Imbalanced Data Stream

Bartosz Przybyl, Jerzy Stefanowski

Learning classifiers from imbalanced and concept drifting data streams is still a challenge. Most of the current proposals focus on taking into account changes in the global imbalance ratio only and ignore the local difficulty factors, such as the minority class decomposition into sub-concepts and the presence of unsafe types of examples (borderline or rare ones). As the above factors present in the stream may deteriorate the performance of popular online classifiers, we propose extensions of resampling online bagging, namely Neighbourhood Undersampling or Oversampling Online Bagging to take better account of the presence of unsafe minority examples. The performed computational experiments with synthetic complex imbalanced data streams have shown their advantage over earlier variants of online bagging resampling ensembles.

comment: 16 pages, 4 figures

☆ Fine-Grained Expressive Power of Weisfeiler-Leman: A Homomorphism Counting Perspective

Junru Zhou, Muhan Zhang

The ability of graph neural networks (GNNs) to count homomorphisms has recently been proposed as a practical and fine-grained measure of their expressive power. Although several existing works have investigated the homomorphism counting power of certain GNN families, a simple and unified framework for analyzing the problem is absent. In this paper, we first propose \emph{generalized folklore Weisfeiler-Leman (GFWL)} algorithms as a flexible design basis for expressive GNNs, and then provide a theoretical framework to algorithmically determine the homomorphism counting power of an arbitrary class of GNN within the GFWL design space. As the considered design space is large enough to accommodate almost all known powerful GNNs, our result greatly extends all existing works, and may find its application in the automation of GNN model design.

☆ Stabilized Neural Prediction of Potential Outcomes in Continuous Time

Konstantin Hess, Stefan Feuerriegel

Patient trajectories from electronic health records are widely used to predict potential outcomes of treatments over time, which then allows to personalize care. Yet, existing neural methods for this purpose have a key limitation: while some adjust for time-varying confounding, these methods assume that the time series are recorded in discrete time. In other words, they are constrained to settings where measurements and treatments are conducted at fixed time steps, even though this is unrealistic in medical practice. In this work, we aim to predict potential outcomes in continuous time. The latter is of direct practical relevance because it allows for modeling patient trajectories where measurements and treatments take place at arbitrary, irregular timestamps. We thus propose a new method called stabilized continuous time inverse propensity network (SCIP-Net). For this, we further derive stabilized inverse propensity weights for robust prediction of the potential outcomes. To the best of our knowledge, our SCIP-Net is the first neural method that performs proper adjustments for time-varying confounding in continuous time.

☆ Authentication by Location Tracking in Underwater Acoustic Networks

Gianmaria Ventura, Francesco Ardizzon, Stefano Tomasin

Physical layer message authentication in underwater acoustic networks (UWANs) leverages the characteristics of the underwater acoustic channel (UWAC) as a fingerprint of the transmitting device. However, as the device moves its UWAC changes, and the authentication mechanism must track such variations. In this paper, we propose a context-based authentication mechanism operating in two steps: first, we estimate the position of the underwater device, then we predict its future position based on the previously estimated ones. To check the authenticity of the transmission, we compare the estimated and the predicted position. The location is estimated using a convolutional neural network taking as input the sample covariance matrix of the estimated UWACs. The prediction uses either a Kalman filter or a recurrent neural network (RNN). The authentication check is performed on the squared error between the predicted and estimated positions. The solution based on the Kalman filter outperforms that built on the RNN when the device moves according to a correlated Gauss-Markov mobility model, which reproduces a typical underwater motion.

comment: Article submitted to IEEE Transaction on Wireless Communications

☆ Classification-Denoising Networks

Louis Thiry, Florentin Guth

comment: 18 pages, 5 figures

☆ FedStein: Enhancing Multi-Domain Federated Learning Through James-Stein Estimator NeurIPS 2024

Sunny Gupta, Nikita Jangid, Amit Sethi

comment: 12 pages, 2 figures. Accepted at International Workshop on Federated Foundation Models In Conjunction with NeurIPS 2024 (FL@FM-NeurIPS'24)

☆ Collaborative and Efficient Personalization with Mixtures of Adaptors

Abdulla Jasem Almansoori, Samuel Horváth, Martin Takáč

Non-iid data is prevalent in real-world federated learning problems. Data heterogeneity can come in different types in terms of distribution shifts. In this work, we are interested in the heterogeneity that comes from concept shifts, i.e., shifts in the prediction across clients. In particular, we consider multi-task learning, where we want the model to adapt to the task of the client. We propose a parameter-efficient framework to tackle this issue, where each client learns to mix between parameter-efficient adaptors according to its task. We use Low-Rank Adaptors (LoRAs) as the backbone and extend its concept to other types of layers. We call our framework Federated Low-Rank Adaptive Learning (FLoRAL). This framework is not an algorithm but rather a model parameterization for a multi-task learning objective, so it can work on top of any algorithm that optimizes this objective, which includes many algorithms from the literature. FLoRAL is memory-efficient, and clients are personalized with small states (e.g., one number per adaptor) as the adaptors themselves are federated. Hence, personalization is--in this sense--federated as well. Even though clients can personalize more freely by training an adaptor locally, we show that collaborative and efficient training of adaptors is possible and performs better. We also show that FLoRAL can outperform an ensemble of full models with optimal cluster assignment, which demonstrates the benefits of federated personalization and the robustness of FLoRAL to overfitting. We show promising experimental results on synthetic datasets, real-world federated multi-task problems such as MNIST, CIFAR-10, and CIFAR-100. We also provide a theoretical analysis of local SGD on a relaxed objective and discuss the effects of aggregation mismatch on convergence.

comment: 36 pages, 10 figures

☆ Fourier PINNs: From Strong Boundary Conditions to Adaptive Fourier Bases

Madison Cooley, Varun Shankar, Robert M. Kirby, Shandian Zhe

Interest is rising in Physics-Informed Neural Networks (PINNs) as a mesh-free alternative to traditional numerical solvers for partial differential equations (PDEs). However, PINNs often struggle to learn high-frequency and multi-scale target solutions. To tackle this problem, we first study a strong Boundary Condition (BC) version of PINNs for Dirichlet BCs and observe a consistent decline in relative error compared to the standard PINNs. We then perform a theoretical analysis based on the Fourier transform and convolution theorem. We find that strong BC PINNs can better learn the amplitudes of high-frequency components of the target solutions. However, constructing the architecture for strong BC PINNs is difficult for many BCs and domain geometries. Enlightened by our theoretical analysis, we propose Fourier PINNs -- a simple, general, yet powerful method that augments PINNs with pre-specified, dense Fourier bases. Our proposed architecture likewise learns high-frequency components better but places no restrictions on the particular BCs or problem domains. We develop an adaptive learning and basis selection algorithm via alternating neural net basis optimization, Fourier and neural net basis coefficient estimation, and coefficient truncation. This scheme can flexibly identify the significant frequencies while weakening the nominal frequencies to better capture the target solution's power spectrum. We show the advantage of our approach through a set of systematic experiments.

comment: 24 pages, 15 figures

☆ Generative Artificial Intelligence for Navigating Synthesizable Chemical Space

Wenhao Gao, Shitong Luo, Connor W. Coley

We introduce SynFormer, a generative modeling framework designed to efficiently explore and navigate synthesizable chemical space. Unlike traditional molecular generation approaches, we generate synthetic pathways for molecules to ensure that designs are synthetically tractable. By incorporating a scalable transformer architecture and a diffusion module for building block selection, SynFormer surpasses existing models in synthesizable molecular design. We demonstrate SynFormer's effectiveness in two key applications: (1) local chemical space exploration, where the model generates synthesizable analogs of a reference molecule, and (2) global chemical space exploration, where the model aims to identify optimal molecules according to a black-box property prediction oracle. Additionally, we demonstrate the scalability of our approach via the improvement in performance as more computational resources become available. With our code and trained models openly available, we hope that SynFormer will find use across applications in drug discovery and materials science.

☆ A Multimodal Framework for Deepfake Detection

Kashish Gandhi, Prutha Kulkarni, Taran Shah, Piyush Chaudhari, Meera Narvekar, Kranti Ghag

comment: 22 pages, 14 figures, Accepted in Journal of Electrical Systems

☆ VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning

Han Lin, Tushar Nagarajan, Nicolas Ballas, Mido Assran, Mojtaba Komeili, Mohit Bansal, Koustuv Sinha

comment: 10 pages

☆ On the Hardness of Learning One Hidden Layer Neural Networks

Shuchen Li, Ilias Zadik, Manolis Zampetakis

In this work, we consider the problem of learning one hidden layer ReLU neural networks with inputs from $\mathbb{R}^d$. We show that this learning problem is hard under standard cryptographic assumptions even when: (1) the size of the neural network is polynomial in $d$, (2) its input distribution is a standard Gaussian, and (3) the noise is Gaussian and polynomially small in $d$. Our hardness result is based on the hardness of the Continuous Learning with Errors (CLWE) problem, and in particular, is based on the largely believed worst-case hardness of approximately solving the shortest vector problem up to a multiplicative polynomial factor.

comment: 18 pages

☆ Vulnerability Detection via Topological Analysis of Attention Maps

Pavel Snopov, Andrey Nikolaevich Golubinskiy

Recently, deep learning (DL) approaches to vulnerability detection have gained significant traction. These methods demonstrate promising results, often surpassing traditional static code analysis tools in effectiveness. In this study, we explore a novel approach to vulnerability detection utilizing the tools from topological data analysis (TDA) on the attention matrices of the BERT model. Our findings reveal that traditional machine learning (ML) techniques, when trained on the topological features extracted from these attention matrices, can perform competitively with pre-trained language models (LLMs) such as CodeBERTa. This suggests that TDA tools, including persistent homology, are capable of effectively capturing semantic information critical for identifying vulnerabilities.

comment: Accepted to ITaS2024. Contains 8 pages

☆ S7: Selective and Simplified State Space Layers for Sequence Modeling

Taylan Soydan, Nikola Zubić, Nico Messikommer, Siddhartha Mishra, Davide Scaramuzza

A central challenge in sequence modeling is efficiently handling tasks with extended contexts. While recent state-space models (SSMs) have made significant progress in this area, they often lack input-dependent filtering or require substantial increases in model complexity to handle input variability. We address this gap by introducing S7, a simplified yet powerful SSM that can handle input dependence while incorporating stable reparameterization and specific design choices to dynamically adjust state transitions based on input content, maintaining efficiency and performance. We prove that this reparameterization ensures stability in long-sequence modeling by keeping state transitions well-behaved over time. Additionally, it controls the gradient norm, enabling efficient training and preventing issues like exploding or vanishing gradients. S7 significantly outperforms baselines across various sequence modeling tasks, including neuromorphic event-based datasets, Long Range Arena benchmarks, and various physical and biological time series. Overall, S7 offers a more straightforward approach to sequence modeling without relying on complex, domain-specific inductive biases, achieving significant improvements across key benchmarks.

comment: 23 pages, 3 figures, 11 tables. Equal contribution by Taylan Soydan and Nikola Zubi\'c

☆ Diffusion State-Guided Projected Gradient for Inverse Problems

Rayhan Zirvi, Bahareh Tolooshams, Anima Anandkumar

comment: preprint. under review. RZ and BT have equal contributions

☆ Linear Transformer Topological Masking with Graph Random Features

Isaac Reid, Kumar Avinava Dubey, Deepali Jain, Will Whitney, Amr Ahmed, Joshua Ainslie, Alex Bewley, Mithun Jacob, Aranyak Mehta, David Rendleman, Connor Schenck, Richard E. Turner, René Wagner, Adrian Weller, Krzysztof Choromanski

When training transformers on graph-structured data, incorporating information about the underlying topology is crucial for good performance. Topological masking, a type of relative position encoding, achieves this by upweighting or downweighting attention depending on the relationship between the query and keys in a graph. In this paper, we propose to parameterise topological masks as a learnable function of a weighted adjacency matrix -- a novel, flexible approach which incorporates a strong structural inductive bias. By approximating this mask with graph random features (for which we prove the first known concentration bounds), we show how this can be made fully compatible with linear attention, preserving $\mathcal{O}(N)$ time and space complexity with respect to the number of input tokens. The fastest previous alternative was $\mathcal{O}(N \log N)$ and only suitable for specific graphs. Our efficient masking algorithms provide strong performance gains for tasks on image and point cloud data, including with $>30$k nodes.

☆ Auto-GDA: Automatic Domain Adaptation for Efficient Grounding Verification in Retrieval Augmented Generation

Tobias Leemann, Periklis Petridis, Giuseppe Vietri, Dionysis Manousakas, Aaron Roth, Sergul Aydore

☆ Generative Semantic Communication for Text-to-Speech Synthesis

Jiahao Zheng, Jinke Ren, Peng Xu, Zhihao Yuan, Jie Xu, Fangxin Wang, Gui Gui, Shuguang Cui

Semantic communication is a promising technology to improve communication efficiency by transmitting only the semantic information of the source data. However, traditional semantic communication methods primarily focus on data reconstruction tasks, which may not be efficient for emerging generative tasks such as text-to-speech (TTS) synthesis. To address this limitation, this paper develops a novel generative semantic communication framework for TTS synthesis, leveraging generative artificial intelligence technologies. Firstly, we utilize a pre-trained large speech model called WavLM and the residual vector quantization method to construct two semantic knowledge bases (KBs) at the transmitter and receiver, respectively. The KB at the transmitter enables effective semantic extraction, while the KB at the receiver facilitates lifelike speech synthesis. Then, we employ a transformer encoder and a diffusion model to achieve efficient semantic coding without introducing significant communication overhead. Finally, numerical results demonstrate that our framework achieves much higher fidelity for the generated speech than four baselines, in both cases with additive white Gaussian noise channel and Rayleigh fading channel.

comment: The paper has been accepted by IEEE Globecom Workshop

☆ MLLM as Retriever: Interactively Learning Multimodal Retrieval for Embodied Agents

Junpeng Yue, Xinru Xu, Börje F. Karlsson, Zongqing Lu

MLLM agents demonstrate potential for complex embodied tasks by retrieving multimodal task-relevant trajectory data. However, current retrieval methods primarily focus on surface-level similarities of textual or visual cues in trajectories, neglecting their effectiveness for the specific task at hand. To address this issue, we propose a novel method, MLLM as ReTriever (MART), which enhances the performance of embodied agents by utilizing interaction data to fine-tune an MLLM retriever based on preference learning, such that the retriever fully considers the effectiveness of trajectories and prioritize them for unseen tasks. We also introduce Trajectory Abstraction, a mechanism that leverages MLLMs' summarization capabilities to represent trajectories with fewer tokens while preserving key information, enabling agents to better comprehend milestones in the trajectory. Experimental results across various environments demonstrate our method significantly improves task success rates in unseen scenes compared to baseline methods. This work presents a new paradigm for multimodal retrieval in embodied agents, by fine-tuning a general-purpose MLLM as the retriever to assess trajectory effectiveness. All benchmark task sets and simulator code modifications for action and observation spaces will be released.

☆ Zebra: In-Context and Generative Pretraining for Solving Parametric PDEs

Louis Serrano, Armand Kassaï Koupaï, Thomas X Wang, Pierre Erbacher, Patrick Gallinari

Solving time-dependent parametric partial differential equations (PDEs) is challenging, as models must adapt to variations in parameters such as coefficients, forcing terms, and boundary conditions. Data-driven neural solvers either train on data sampled from the PDE parameters distribution in the hope that the model generalizes to new instances or rely on gradient-based adaptation and meta-learning to implicitly encode the dynamics from observations. This often comes with increased inference complexity. Inspired by the in-context learning capabilities of large language models (LLMs), we introduce Zebra, a novel generative auto-regressive transformer designed to solve parametric PDEs without requiring gradient adaptation at inference. By leveraging in-context information during both pre-training and inference, Zebra dynamically adapts to new tasks by conditioning on input sequences that incorporate context trajectories or preceding states. This approach enables Zebra to flexibly handle arbitrarily sized context inputs and supports uncertainty quantification through the sampling of multiple solution trajectories. We evaluate Zebra across a variety of challenging PDE scenarios, demonstrating its adaptability, robustness, and superior performance compared to existing approaches.

☆ A General Framework for Producing Interpretable Semantic Text Embeddings

Yiqun Sun, Qiang Huang, Yixuan Tang, Anthony K. H. Tung, Jun Yu

comment: 19 pages, 5 figures, and 9 tables

☆ EB-NeRD: A Large-Scale Dataset for News Recommendation RecSys '24

Johannes Kruse, Kasper Lindskow, Saikishore Kalloori, Marco Polignano, Claudio Pomo, Abhishek Srivastava, Anshuk Uppal, Michael Riis Andersen, Jes Frellsen

Personalized content recommendations have been pivotal to the content experience in digital media from video streaming to social networks. However, several domain specific challenges have held back adoption of recommender systems in news publishing. To address these challenges, we introduce the Ekstra Bladet News Recommendation Dataset (EB-NeRD). The dataset encompasses data from over a million unique users and more than 37 million impression logs from Ekstra Bladet. It also includes a collection of over 125,000 Danish news articles, complete with titles, abstracts, bodies, and metadata, such as categories. EB-NeRD served as the benchmark dataset for the RecSys '24 Challenge, where it was demonstrated how the dataset can be used to address both technical and normative challenges in designing effective and responsible recommender systems for news publishing. The dataset is available at: https://recsys.eb.dk.

comment: 11 pages, 8 tables, 2 figures, RecSys '24

☆ Cayley Graph Propagation

JJ Wilson, Maya Bechler-Speicher, Petar Veličković

In spite of the plethora of success stories with graph neural networks (GNNs) on modelling graph-structured data, they are notoriously vulnerable to over-squashing, whereby tasks necessitate the mixing of information between distance pairs of nodes. To address this problem, prior work suggests rewiring the graph structure to improve information flow. Alternatively, a significant body of research has dedicated itself to discovering and precomputing bottleneck-free graph structures to ameliorate over-squashing. One well regarded family of bottleneck-free graphs within the mathematical community are expander graphs, with prior work$\unicode{x2014}$Expander Graph Propagation (EGP)$\unicode{x2014}$proposing the use of a well-known expander graph family$\unicode{x2014}$the Cayley graphs of the $\mathrm{SL}(2,\mathbb{Z}_n)$ special linear group$\unicode{x2014}$as a computational template for GNNs. However, in EGP the computational graphs used are truncated to align with a given input graph. In this work, we show that truncation is detrimental to the coveted expansion properties. Instead, we propose CGP, a method to propagate information over a complete Cayley graph structure, thereby ensuring it is bottleneck-free to better alleviate over-squashing. Our empirical evidence across several real-world datasets not only shows that CGP recovers significant improvements as compared to EGP, but it is also akin to or outperforms computationally complex graph rewiring techniques.

comment: 20 pages, 6 figures

☆ Aircraft Radar Altimeter Interference Mitigation Through a CNN-Layer Only Denoising Autoencoder Architecture

Samuel B. Brown, Stephen Young, Adam Wagenknecht, Daniel Jakubisin, Charles E. Thornton, Aaron Orndorff, William C. Headley

Denoising autoencoders for signal processing applications have been shown to experience significant difficulty in learning to reconstruct radio frequency communication signals, particularly in the large sample regime. In communication systems, this challenge is primarily due to the need to reconstruct the modulated data stream which is generally highly stochastic in nature. In this work, we take advantage of this limitation by using the denoising autoencoder to instead remove interfering radio frequency communication signals while reconstructing highly structured FMCW radar signals. More specifically, in this work we show that a CNN-layer only autoencoder architecture can be utilized to improve the accuracy of a radar altimeter's ranging estimate even in severe interference environments consisting of a multitude of interference signals. This is demonstrated through comprehensive performance analysis of an end-to-end FMCW radar altimeter simulation with and without the convolutional layer-only autoencoder. The proposed approach significantly improves interference mitigation in the presence of both narrow-band tone interference as well as wideband QPSK interference in terms of range RMS error, number of false altitude reports, and the peak-to-sidelobe ratio of the resulting range profile. FMCW radar signals of up to 40,000 IQ samples can be reliably reconstructed.

comment: To be presented at MILCOM 2024, Washington DC

☆ Benchmarking the Fidelity and Utility of Synthetic Relational Data

Valter Hudovernik, Martin Jurkovič, Erik Štrumbelj

Synthesizing relational data has started to receive more attention from researchers, practitioners, and industry. The task is more difficult than synthesizing a single table due to the added complexity of relationships between tables. For the same reason, benchmarking methods for synthesizing relational data introduces new challenges. Our work is motivated by a lack of an empirical evaluation of state-of-the-art methods and by gaps in the understanding of how such an evaluation should be done. We review related work on relational data synthesis, common benchmarking datasets, and approaches to measuring the fidelity and utility of synthetic data. We combine the best practices and a novel robust detection approach into a benchmarking tool and use it to compare six methods, including two commercial tools. While some methods are better than others, no method is able to synthesize a dataset that is indistinguishable from original data. For utility, we typically observe moderate correlation between real and synthetic data for both model predictive performance and feature importance.

☆ Predictive Coding for Decision Transformer IROS 2024

Tung M. Luu, Donghoon Lee, Chang D. Yoo

Recent work in offline reinforcement learning (RL) has demonstrated the effectiveness of formulating decision-making as return-conditioned supervised learning. Notably, the decision transformer (DT) architecture has shown promise across various domains. However, despite its initial success, DTs have underperformed on several challenging datasets in goal-conditioned RL. This limitation stems from the inefficiency of return conditioning for guiding policy learning, particularly in unstructured and suboptimal datasets, resulting in DTs failing to effectively learn temporal compositionality. Moreover, this problem might be further exacerbated in long-horizon sparse-reward tasks. To address this challenge, we propose the Predictive Coding for Decision Transformer (PCDT) framework, which leverages generalized future conditioning to enhance DT methods. PCDT utilizes an architecture that extends the DT framework, conditioned on predictive codings, enabling decision-making based on both past and future factors, thereby improving generalization. Through extensive experiments on eight datasets from the AntMaze and FrankaKitchen environments, our proposed method achieves performance on par with or surpassing existing popular value-based and transformer-based methods in offline goal-conditioned RL. Furthermore, we also evaluate our method on a goal-reaching task with a physical robot.

comment: 8 pages, IROS 2024 (Code: https://github.com/tunglm2203/pcdt)

☆ Conformal confidence sets for biomedical image segmentation

Samuel Davenport

We develop confidence sets which provide spatial uncertainty guarantees for the output of a black-box machine learning model designed for image segmentation. To do so we adapt conformal inference to the imaging setting, obtaining thresholds on a calibration dataset based on the distribution of the maximum of the transformed logit scores within and outside of the ground truth masks. We prove that these confidence sets, when applied to new predictions of the model, are guaranteed to contain the true unknown segmented mask with desired probability. We show that learning appropriate score transformations on a learning dataset before performing calibration is crucial for optimizing performance. We illustrate and validate our approach on a polpys tumor dataset. To do so we obtain the logit scores from a deep neural network trained for polpys segmentation and show that using distance transformed scores to obtain outer confidence sets and the original scores for inner confidence sets enables tight bounds on tumor location whilst controlling the false coverage rate.

☆ Distributed Networked Multi-task Learning

Lingzhou Hong, Alfredo Garcia

We consider a distributed multi-task learning scheme that accounts for multiple linear model estimation tasks with heterogeneous and/or correlated data streams. We assume that nodes can be partitioned into groups corresponding to different learning tasks and communicate according to a directed network topology. Each node estimates a linear model asynchronously and is subject to local (within-group) regularization and global (across groups) regularization terms targeting noise reduction and generalization performance improvement respectively. We provide a finite-time characterization of convergence of the estimators and task relation and illustrate the scheme's general applicability in two examples: random field temperature estimation and modeling student performance from different academic districts.

☆ EBES: Easy Benchmarking for Event Sequences

Dmitry Osin, Igor Udovichenko, Viktor Moskvoretskii, Egor Shvetsov, Evgeny Burnaev

Event sequences, characterized by irregular sampling intervals and a mix of categorical and numerical features, are common data structures in various real-world domains such as healthcare, finance, and user interaction logs. Despite advances in temporal data modeling techniques, there is no standardized benchmarks for evaluating their performance on event sequences. This complicates result comparison across different papers due to varying evaluation protocols, potentially misleading progress in this field. We introduce EBES, a comprehensive benchmarking tool with standardized evaluation scenarios and protocols, focusing on regression and classification problems with sequence-level targets. Our library simplifies benchmarking, dataset addition, and method integration through a unified interface. It includes a novel synthetic dataset and provides preprocessed real-world datasets, including the largest publicly available banking dataset. Our results provide an in-depth analysis of datasets, identifying some as unsuitable for model comparison. We investigate the importance of modeling temporal and sequential components, as well as the robustness and scaling properties of the models. These findings highlight potential directions for future research. Our benchmark aim is to facilitate reproducible research, expediting progress and increasing real-world impacts.

☆ GraphCroc: Cross-Correlation Autoencoder for Graph Structural Reconstruction NeurIPS 2024

Shijin Duan, Ruyi Ding, Jiaxing He, Aidong Adam Ding, Yunsi Fei, Xiaolin Xu

Graph-structured data is integral to many applications, prompting the development of various graph representation methods. Graph autoencoders (GAEs), in particular, reconstruct graph structures from node embeddings. Current GAE models primarily utilize self-correlation to represent graph structures and focus on node-level tasks, often overlooking multi-graph scenarios. Our theoretical analysis indicates that self-correlation generally falls short in accurately representing specific graph features such as islands, symmetrical structures, and directional edges, particularly in smaller or multiple graph contexts. To address these limitations, we introduce a cross-correlation mechanism that significantly enhances the GAE representational capabilities. Additionally, we propose GraphCroc, a new GAE that supports flexible encoder architectures tailored for various downstream tasks and ensures robust structural reconstruction, through a mirrored encoding-decoding process. This model also tackles the challenge of representation bias during optimization by implementing a loss-balancing strategy. Both theoretical analysis and numerical evaluations demonstrate that our methodology significantly outperforms existing self-correlation-based GAEs in graph structure reconstruction.

comment: 22 pages, 16 figures. Accepted in NeurIPS 2024

☆ Lightning UQ Box: A Comprehensive Framework for Uncertainty Quantification in Deep Learning

Nils Lehmann, Jakob Gawlikowski, Adam J. Stewart, Vytautas Jancauskas, Stefan Depeweg, Eric Nalisnick, Nina Maria Gottschling

comment: 10 pages, 8 figures

☆ From Epilepsy Seizures Classification to Detection: A Deep Learning-based Approach for Raw EEG Signals

Davy Darankoum, Manon Villalba, Clelia Allioux, Baptiste Caraballo, Carine Dumont, Eloise Gronlier, Corinne Roucard, Yann Roche, Chloe Habermacher, Sergei Grudinin, Julien Volle

Epilepsy represents the most prevalent neurological disease in the world. One-third of people suffering from mesial temporal lobe epilepsy (MTLE) exhibit drug resistance, urging the need to develop new treatments. A key part in anti-seizure medication (ASM) development is the capability of detecting and quantifying epileptic seizures occurring in electroencephalogram (EEG) signals, which is crucial for treatment efficacy evaluation. In this study, we introduced a seizure detection pipeline based on deep learning models applied to raw EEG signals. This pipeline integrates: a new pre-processing technique which segments continuous raw EEG signals without prior distinction between seizure and seizure-free activities; a post-processing algorithm developed to reassemble EEG segments and allow the identification of seizures start/end; and finally, a new evaluation procedure based on a strict seizure events comparison between predicted and real labels. Models training have been performed using a data splitting strategy which addresses the potential for data leakage. We demonstrated the fundamental differences between a seizure classification and a seizure detection task and showed the differences in performance between the two tasks. Finally, we demonstrated the generalization capabilities across species of our best architecture, combining a Convolutional Neural Network and a Transformer encoder. The model was trained on animal EEGs and tested on human EEGs with a F1-score of 93% on a balanced Bonn dataset.

comment: 25 pages, 7 tables, 4 figures

☆ Predicting perturbation targets with causal differential networks

Menghua Wu, Umesh Padia, Sean H. Murphy, Regina Barzilay, Tommi Jaakkola

Rationally identifying variables responsible for changes to a biological system can enable myriad applications in disease understanding and cell engineering. From a causality perspective, we are given two datasets generated by the same causal model, one observational (control) and one interventional (perturbed). The goal is to isolate the subset of measured variables (e.g. genes) that were the targets of the intervention, i.e. those whose conditional independencies have changed. Knowing the causal graph would limit the search space, allowing us to efficiently pinpoint these variables. However, current algorithms that infer causal graphs in the presence of unknown intervention targets scale poorly to the hundreds or thousands of variables in biological data, as they must jointly search the combinatorial spaces of graphs and consistent intervention targets. In this work, we propose a causality-inspired approach for predicting perturbation targets that decouples the two search steps. First, we use an amortized causal discovery model to separately infer causal graphs from the observational and interventional datasets. Then, we learn to map these paired graphs to the sets of variables that were intervened upon, in a supervised learning framework. This approach consistently outperforms baselines for perturbation modeling on seven single-cell transcriptomics datasets, each with thousands of measured variables. We also demonstrate significant improvements over six causal discovery algorithms in predicting intervention targets across a variety of tractable, synthetic datasets.

☆ Mitigating Adversarial Perturbations for Deep Reinforcement Learning via Vector Quantization IROS 2024

Tung M. Luu, Thanh Nguyen, Tee Joshua Tian Jin, Sungwoon Kim, Chang D. Yoo

Recent studies reveal that well-performing reinforcement learning (RL) agents in training often lack resilience against adversarial perturbations during deployment. This highlights the importance of building a robust agent before deploying it in the real world. Most prior works focus on developing robust training-based procedures to tackle this problem, including enhancing the robustness of the deep neural network component itself or adversarially training the agent on strong attacks. In this work, we instead study an input transformation-based defense for RL. Specifically, we propose using a variant of vector quantization (VQ) as a transformation for input observations, which is then used to reduce the space of adversarial attacks during testing, resulting in the transformed observations being less affected by attacks. Our method is computationally efficient and seamlessly integrates with adversarial training, further enhancing the robustness of RL agents against adversarial attacks. Through extensive experiments in multiple environments, we demonstrate that using VQ as the input transformation effectively defends against adversarial attacks on the agent's observations.

comment: 8 pages, IROS 2024 (Code: https://github.com/tunglm2203/vq_robust_rl)

☆ Make Interval Bound Propagation great again

Patryk Krukowski, Daniel Wilczak, Jacek Tabor, Anna Bielawska, Przemysław Spurek

In various scenarios motivated by real life, such as medical data analysis, autonomous driving, and adversarial training, we are interested in robust deep networks. A network is robust when a relatively small perturbation of the input cannot lead to drastic changes in output (like change of class, etc.). This falls under the broader scope field of Neural Network Certification (NNC). Two crucial problems in NNC are of profound interest to the scientific community: how to calculate the robustness of a given pre-trained network and how to construct robust networks. The common approach to constructing robust networks is Interval Bound Propagation (IBP). This paper demonstrates that IBP is sub-optimal in the first case due to its susceptibility to the wrapping effect. Even for linear activation, IBP gives strongly sub-optimal bounds. Consequently, one should use strategies immune to the wrapping effect to obtain bounds close to optimal ones. We adapt two classical approaches dedicated to strict computations -- Dubleton Arithmetic and Affine Arithmetic -- to mitigate the wrapping effect in neural networks. These techniques yield precise results for networks with linear activation functions, thus resisting the wrapping effect. As a result, we achieve bounds significantly closer to the optimal level than IBPs.

☆ Latent Abstractions in Generative Diffusion Models

Giulio Franzese, Mattia Martini, Giulio Corallo, Paolo Papotti, Pietro Michiardi

In this work we study how diffusion-based generative models produce high-dimensional data, such as an image, by implicitly relying on a manifestation of a low-dimensional set of latent abstractions, that guide the generative process. We present a novel theoretical framework that extends NLF, and that offers a unique perspective on SDE-based generative models. The development of our theory relies on a novel formulation of the joint (state and measurement) dynamics, and an information-theoretic measure of the influence of the system state on the measurement process. According to our theory, diffusion models can be cast as a system of SDE, describing a non-linear filter in which the evolution of unobservable latent abstractions steers the dynamics of an observable measurement process (corresponding to the generative pathways). In addition, we present an empirical study to validate our theory and previous empirical results on the emergence of latent abstractions at different stages of the generative process.

☆ Error Correction Code Transformer: From Non-Unified to Unified

Yongli Yan, Jieao Zhu, Tianyue Zheng, Jiaqi He, Linglong Dai

Channel coding is vital for reliable data transmission in modern wireless systems, and its significance will increase with the emergence of sixth-generation (6G) networks, which will need to support various error correction codes. However, traditional decoders were typically designed as fixed hardware circuits tailored to specific decoding algorithms, leading to inefficiencies and limited flexibility. To address these challenges, this paper proposes a unified, code-agnostic Transformer-based decoding architecture capable of handling multiple linear block codes, including Polar, Low-Density Parity-Check (LDPC), and Bose-Chaudhuri-Hocquenghem (BCH), within a single framework. To achieve this, standardized units are employed to harmonize parameters across different code types, while the redesigned unified attention module compresses the structural information of various codewords. Additionally, a sparse mask, derived from the sparsity of the parity-check matrix, is introduced to enhance the model's ability to capture inherent constraints between information and parity-check bits, resulting in improved decoding accuracy and robustness. Extensive experimental results demonstrate that the proposed unified Transformer-based decoder not only outperforms existing methods but also provides a flexible, efficient, and high-performance solution for next-generation wireless communication systems.

☆ Dolphin: A Programmable Framework for Scalable Neurosymbolic Learning

Aaditya Naik, Jason Liu, Claire Wang, Saikat Dutta, Mayur Naik, Eric Wong

Neurosymbolic learning has emerged as a promising paradigm to incorporate symbolic reasoning into deep learning models. However, existing frameworks are limited in scalability with respect to both the training data and the complexity of symbolic programs. We propose Dolphin, a framework to scale neurosymbolic learning at a fundamental level by mapping both forward chaining and backward gradient propagation in symbolic programs to vectorized computations. For this purpose, Dolphin introduces a set of abstractions and primitives built directly on top of a high-performance deep learning framework like PyTorch, effectively enabling symbolic programs to be written as PyTorch modules. It thereby enables neurosymbolic programs to be written in a language like Python that is familiar to developers and compile them to computation graphs that are amenable to end-to-end differentiation on GPUs. We evaluate Dolphin on a suite of 13 benchmarks across 5 neurosymbolic tasks that combine deep learning models for text, image, or video processing with symbolic programs that involve multi-hop reasoning, recursion, and even black-box functions like Python eval(). Dolphin only takes 0.33%-37.17% of the time (and 2.77% on average) to train these models on the largest input per task compared to baselines Scallop, ISED, and IndeCateR+, which time out on most of these inputs. Models written in Dolphin also achieve state-of-the-art accuracies even on the largest benchmarks.

☆ Audio-Agent: Leveraging LLMs For Audio Generation, Editing and Composition

Zixuan Wang, Yu-Wing Tai, Chi-Keung Tang

☆ Influence-oriented Personalized Federated Learning

Yue Tan, Guodong Long, Jing Jiang, Chengqi Zhang

Traditional federated learning (FL) methods often rely on fixed weighting for parameter aggregation, neglecting the mutual influence by others. Hence, their effectiveness in heterogeneous data contexts is limited. To address this problem, we propose an influence-oriented federated learning framework, namely FedC^2I, which quantitatively measures Client-level and Class-level Influence to realize adaptive parameter aggregation for each client. Our core idea is to explicitly model the inter-client influence within an FL system via the well-crafted influence vector and influence matrix. The influence vector quantifies client-level influence, enables clients to selectively acquire knowledge from others, and guides the aggregation of feature representation layers. Meanwhile, the influence matrix captures class-level influence in a more fine-grained manner to achieve personalized classifier aggregation. We evaluate the performance of FedC^2I against existing federated learning methods under non-IID settings and the results demonstrate the superiority of our method.

☆ Quo Vadis, Motion Generation? From Large Language Models to Large Motion Models

Ye Wang, Sipeng Zheng, Bin Cao, Qianshan Wei, Qin Jin, Zongqing Lu

☆ Selective Test-Time Adaptation for Unsupervised Anomaly Detection using Neural Implicit Representations MICCAI

Sameer Ambekar, Julia A. Schnabel, Cosmin Bereca

Deep learning models in medical imaging often encounter challenges when adapting to new clinical settings unseen during training. Test-time adaptation offers a promising approach to optimize models for these unseen domains, yet its application in anomaly detection (AD) remains largely unexplored. AD aims to efficiently identify deviations from normative distributions; however, full adaptation, including pathological shifts, may inadvertently learn the anomalies it intends to detect. We introduce a novel concept of \emph{selective} test-time adaptation that utilizes the inherent characteristics of deep pre-trained features to adapt \emph{selectively} in a zero-shot manner to any test image from an unseen domain. This approach employs a model-agnostic, lightweight multi-layer perceptron for neural implicit representations, enabling the adaptation of outputs from any reconstruction-based AD method without altering the source-trained model. Rigorous validation in brain AD demonstrated that our strategy substantially enhances detection accuracy for multiple conditions and different target distributions. Specifically, our method improves the detection rates by up to 78\% for enlarged ventricles and 24\% for edemas.

comment: Accepted at MICCAIw ADSMI

☆ SELU: Self-Learning Embodied MLLMs in Unknown Environments

Boyu Li, Haobin Jiang, Ziluo Ding, Xinrun Xu, Haoran Li, Dongbin Zhao, Zongqing Lu

☆ Resource-aware Mixed-precision Quantization for Enhancing Deployability of Transformers for Time-series Forecasting on Embedded FPGAs

Tianheng Ling, Chao Qian, Gregor Schiele

This study addresses the deployment challenges of integer-only quantized Transformers on resource-constrained embedded FPGAs (Xilinx Spartan-7 XC7S15). We enhanced the flexibility of our VHDL template by introducing a selectable resource type for storing intermediate results across model layers, thereby breaking the deployment bottleneck by utilizing BRAM efficiently. Moreover, we developed a resource-aware mixed-precision quantization approach that enables researchers to explore hardware-level quantization strategies without requiring extensive expertise in Neural Architecture Search. This method provides accurate resource utilization estimates with a precision discrepancy as low as 3%, compared to actual deployment metrics. Compared to previous work, our approach has successfully facilitated the deployment of model configurations utilizing mixed-precision quantization, thus overcoming the limitations inherent in five previously non-deployable configurations with uniform quantization bitwidths. Consequently, this research enhances the applicability of Transformers in embedded systems, facilitating a broader range of Transformer-powered applications on edge devices.

comment: Accepted by the 21st EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous2024). 20 pages, 8 figures, 6 tables

☆ Five Years of COVID-19 Discourse on Instagram: A Labeled Instagram Dataset of Over Half a Million Posts for Multilingual Sentiment Analysis

Nirmalya Thakur

☆ Demystifying the Token Dynamics of Deep Selective State Space Models

Thieu N Vo, Tung D. Pham, Xin T. Tong, Tan Minh Nguyen

Selective state space models (SSM), such as Mamba, have gained prominence for their effectiveness in modeling sequential data. Despite their outstanding empirical performance, a comprehensive theoretical understanding of deep selective SSM remains elusive, hindering their further development and adoption for applications that need high fidelity. In this paper, we investigate the dynamical properties of tokens in a pre-trained Mamba model. In particular, we derive the dynamical system governing the continuous-time limit of the Mamba model and characterize the asymptotic behavior of its solutions. In the one-dimensional case, we prove that only one of the following two scenarios happens: either all tokens converge to zero, or all tokens diverge to infinity. We provide criteria based on model parameters to determine when each scenario occurs. For the convergent scenario, we empirically verify that this scenario negatively impacts the model's performance. For the divergent scenario, we prove that different tokens will diverge to infinity at different rates, thereby contributing unequally to the updates during model training. Based on these investigations, we propose two refinements for the model: excluding the convergent scenario and reordering tokens based on their importance scores, both aimed at improving practical performance. Our experimental results validate these refinements, offering insights into enhancing Mamba's effectiveness in real-world applications.

☆ Enhanced Transformer architecture for in-context learning of dynamical systems

Matteo Rufolo, Dario Piga, Gabriele Maroni, Marco Forgione

Recently introduced by some of the authors, the in-context identification paradigm aims at estimating, offline and based on synthetic data, a meta-model that describes the behavior of a whole class of systems. Once trained, this meta-model is fed with an observed input/output sequence (context) generated by a real system to predict its behavior in a zero-shot learning fashion. In this paper, we enhance the original meta-modeling framework through three key innovations: by formulating the learning task within a probabilistic framework; by managing non-contiguous context and query windows; and by adopting recurrent patching to effectively handle long context sequences. The efficacy of these modifications is demonstrated through a numerical example focusing on the Wiener-Hammerstein system class, highlighting the model's enhanced performance and scalability.

☆ uniINF: Best-of-Both-Worlds Algorithm for Parameter-Free Heavy-Tailed MABs

Yu Chen, Jiatai Huang, Yan Dai, Longbo Huang

In this paper, we present a novel algorithm, uniINF, for the Heavy-Tailed Multi-Armed Bandits (HTMAB) problem, demonstrating robustness and adaptability in both stochastic and adversarial environments. Unlike the stochastic MAB setting where loss distributions are stationary with time, our study extends to the adversarial setup, where losses are generated from heavy-tailed distributions that depend on both arms and time. Our novel algorithm `uniINF` enjoys the so-called Best-of-Both-Worlds (BoBW) property, performing optimally in both stochastic and adversarial environments without knowing the exact environment type. Moreover, our algorithm also possesses a Parameter-Free feature, i.e., it operates without the need of knowing the heavy-tail parameters $(\sigma, \alpha)$ a-priori. To be precise, uniINF ensures nearly-optimal regret in both stochastic and adversarial environments, matching the corresponding lower bounds when $(\sigma, \alpha)$ is known (up to logarithmic factors). To our knowledge, uniINF is the first parameter-free algorithm to achieve the BoBW property for the heavy-tailed MAB problem. Technically, we develop innovative techniques to achieve BoBW guarantees for Parameter-Free HTMABs, including a refined analysis for the dynamics of log-barrier, an auto-balancing learning rate scheduling scheme, an adaptive skipping-clipping loss tuning technique, and a stopping-time analysis for logarithmic regret.

☆ Neural Sampling from Boltzmann Densities: Fisher-Rao Curves in the Wasserstein Geometry

Jannis Chemseddine, Christian Wald, Richard Duong, Gabriele Steidl

We deal with the task of sampling from an unnormalized Boltzmann density $\rho_D$ by learning a Boltzmann curve given by energies $f_t$ starting in a simple density $\rho_Z$. First, we examine conditions under which Fisher-Rao flows are absolutely continuous in the Wasserstein geometry. Second, we address specific interpolations $f_t$ and the learning of the related density/velocity pairs $(\rho_t,v_t)$. It was numerically observed that the linear interpolation, which requires only a parametrization of the velocity field $v_t$, suffers from a "teleportation-of-mass" issue. Using tools from the Wasserstein geometry, we give an analytical example, where we can precisely measure the explosion of the velocity field. Inspired by M\'at\'e and Fleuret, who parametrize both $f_t$ and $v_t$, we propose an interpolation which parametrizes only $f_t$ and fixes an appropriate $v_t$. This corresponds to the Wasserstein gradient flow of the Kullback-Leibler divergence related to Langevin dynamics. We demonstrate by numerical examples that our model provides a well-behaved flow field which successfully solves the above sampling task.

☆ BN-SCAFFOLD: controlling the drift of Batch Normalization statistics in Federated Learning

Gonzalo Iñaki Quintana, Laurence Vancamberg, Vincent Jugnon, Mathilde Mougeot, Agnès Desolneux

Federated Learning (FL) is gaining traction as a learning paradigm for training Machine Learning (ML) models in a decentralized way. Batch Normalization (BN) is ubiquitous in Deep Neural Networks (DNN), as it improves convergence and generalization. However, BN has been reported to hinder performance of DNNs in heterogeneous FL. Recently, the FedTAN algorithm has been proposed to mitigate the effect of heterogeneity on BN, by aggregating BN statistics and gradients from all the clients. However, it has a high communication cost, that increases linearly with the depth of the DNN. SCAFFOLD is a variance reduction algorithm, that estimates and corrects the client drift in a communication-efficient manner. Despite its promising results in heterogeneous FL settings, it has been reported to underperform for models with BN. In this work, we seek to revive SCAFFOLD, and more generally variance reduction, as an efficient way of training DNN with BN in heterogeneous FL. We introduce a unified theoretical framework for analyzing the convergence of variance reduction algorithms in the BN-DNN setting, inspired of by the work of Wang et al. 2023, and show that SCAFFOLD is unable to remove the bias introduced by BN. We thus propose the BN-SCAFFOLD algorithm, which extends the client drift correction of SCAFFOLD to BN statistics. We prove convergence using the aforementioned framework and validate the theoretical results with experiments on MNIST and CIFAR-10. BN-SCAFFOLD equals the performance of FedTAN, without its high communication cost, outperforming Federated Averaging (FedAvg), SCAFFOLD, and other FL algorithms designed to mitigate BN heterogeneity.

☆ Manikin-Recorded Cardiopulmonary Sounds Dataset Using Digital Stethoscope

Yasaman Torabi, Shahram Shirani, James P. Reilly

Heart and lung sounds are crucial for healthcare monitoring. Recent improvements in stethoscope technology have made it possible to capture patient sounds with enhanced precision. In this dataset, we used a digital stethoscope to capture both heart and lung sounds, including individual and mixed recordings. To our knowledge, this is the first dataset to offer both separate and mixed cardiorespiratory sounds. The recordings were collected from a clinical manikin, a patient simulator designed to replicate human physiological conditions, generating clean heart and lung sounds at different body locations. This dataset includes both normal sounds and various abnormalities (i.e., murmur, atrial fibrillation, tachycardia, atrioventricular block, third and fourth heart sound, wheezing, crackles, rhonchi, pleural rub, and gurgling sounds). The dataset includes audio recordings of chest examinations performed at different anatomical locations, as determined by specialist nurses. Each recording has been enhanced using frequency filters to highlight specific sound types. This dataset is useful for applications in artificial intelligence, such as automated cardiopulmonary disease detection, sound classification, unsupervised separation techniques, and deep learning algorithms related to audio signal processing.

☆ Sm: enhanced localization in Multiple Instance Learning for medical imaging classification NeurIPS 2024

Francisco M. Castro-Macías, Pablo Morales-Álvarez, Yunan Wu, Rafael Molina, Aggelos K. Katsaggelos

comment: 24 pages, 14 figures, 2024 Conference on Neural Information Processing Systems (NeurIPS 2024)

☆ Optimal Transport for $ε$-Contaminated Credal Sets

Michele Caprio

We provide a version for lower probabilities of Monge's and Kantorovich's optimal transport problems. We show that, when the lower probabilities are the lower envelopes of $\epsilon$-contaminated sets, then our version of Monge's, and a restricted version of our Kantorovich's problems, coincide with their respective classical versions. We also give sufficient conditions for the existence of our version of Kantorovich's optimal plan, and for the two problems to be equivalent. As a byproduct, we show that for $\epsilon$-contaminations the lower probability versions of Monge's and Kantorovich's optimal transport problems need not coincide. The applications of our results to Machine Learning and Artificial Intelligence are also discussed.

☆ Test-time Adaptation for Regression by Subspace Alignment

Kazuki Adachi, Shin'ya Yamaguchi, Atsutoshi Kumagai, Tomoki Hamagami

This paper investigates test-time adaptation (TTA) for regression, where a regression model pre-trained in a source domain is adapted to an unknown target distribution with unlabeled target data. Although regression is one of the fundamental tasks in machine learning, most of the existing TTA methods have classification-specific designs, which assume that models output class-categorical predictions, whereas regression models typically output only single scalar values. To enable TTA for regression, we adopt a feature alignment approach, which aligns the feature distributions between the source and target domains to mitigate the domain gap. However, we found that naive feature alignment employed in existing TTA methods for classification is ineffective or even worse for regression because the features are distributed in a small subspace and many of the raw feature dimensions have little significance to the output. For an effective feature alignment in TTA for regression, we propose Significant-subspace Alignment (SSA). SSA consists of two components: subspace detection and dimension weighting. Subspace detection finds the feature subspace that is representative and significant to the output. Then, the feature alignment is performed in the subspace during TTA. Meanwhile, dimension weighting raises the importance of the dimensions of the feature subspace that have greater significance to the output. We experimentally show that SSA outperforms various baselines on real-world datasets.

☆ How much can we forget about Data Contamination?

Sebastian Bordt, Suraj Srinivas, Valentyn Boreiko, Ulrike von Luxburg

☆ Elucidating the Design Choice of Probability Paths in Flow Matching for Forecasting

Soon Hoe Lim, Yijin Wang, Annan Yu, Emma Hart, Michael W. Mahoney, Xiaoye S. Li, N. Benjamin Erichson

Flow matching has recently emerged as a powerful paradigm for generative modeling and has been extended to probabilistic time series forecasting in latent spaces. However, the impact of the specific choice of probability path model on forecasting performance remains under-explored. In this work, we demonstrate that forecasting spatio-temporal data with flow matching is highly sensitive to the selection of the probability path model. Motivated by this insight, we propose a novel probability path model designed to improve forecasting performance. Our empirical results across various dynamical system benchmarks show that our model achieves faster convergence during training and improved predictive performance compared to existing probability path models. Importantly, our approach is efficient during inference, requiring only a few sampling steps. This makes our proposed model practical for real-world applications and opens new avenues for probabilistic forecasting.

comment: 30 pages

♻ ☆ EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage

Zeyi Liao, Lingbo Mo, Chejian Xu, Mintong Kang, Jiawei Zhang, Chaowei Xiao, Yuan Tian, Bo Li, Huan Sun

comment: 29 pages

♻ ☆ Diffusion Models are Evolutionary Algorithms

Yanbo Zhang, Benedikt Hartl, Hananel Hazan, Michael Levin

In a convergence of machine learning and biology, we reveal that diffusion models are evolutionary algorithms. By considering evolution as a denoising process and reversed evolution as diffusion, we mathematically demonstrate that diffusion models inherently perform evolutionary algorithms, naturally encompassing selection, mutation, and reproductive isolation. Building on this equivalence, we propose the Diffusion Evolution method: an evolutionary algorithm utilizing iterative denoising -- as originally introduced in the context of diffusion models -- to heuristically refine solutions in parameter spaces. Unlike traditional approaches, Diffusion Evolution efficiently identifies multiple optimal solutions and outperforms prominent mainstream evolutionary algorithms. Furthermore, leveraging advanced concepts from diffusion models, namely latent space diffusion and accelerated sampling, we introduce Latent Space Diffusion Evolution, which finds solutions for evolutionary tasks in high-dimensional complex parameter space while significantly reducing computational steps. This parallel between diffusion and evolution not only bridges two different fields but also opens new avenues for mutual enhancement, raising questions about open-ended evolution and potentially utilizing non-Gaussian or discrete diffusion models in the context of Diffusion Evolution.

comment: 15 pages, 4 figures, 2 tables

♻ ☆ Searching for Efficient Linear Layers over a Continuous Space of Structured Matrices NeurIPS 2024

Andres Potapczynski, Shikai Qiu, Marc Finzi, Christopher Ferri, Zixi Chen, Micah Goldblum, Bayan Bruss, Christopher De Sa, Andrew Gordon Wilson

Dense linear layers are the dominant computational bottleneck in large neural networks, presenting a critical need for more efficient alternatives. Previous efforts focused on a small number of hand-crafted structured matrices and neglected to investigate whether these structures can surpass dense layers in terms of compute-optimal scaling laws when both the model size and training examples are optimally allocated. In this work, we present a unifying framework that enables searching among all linear operators expressible via an Einstein summation. This framework encompasses many previously proposed structures, such as low-rank, Kronecker, Tensor-Train, Block Tensor-Train (BTT), and Monarch, along with many novel structures. To analyze the framework, we develop a taxonomy of all such operators based on their computational and algebraic properties and show that differences in the compute-optimal scaling laws are mostly governed by a small number of variables that we introduce. Namely, a small $\omega$ (which measures parameter sharing) and large $\psi$ (which measures the rank) reliably led to better scaling laws. Guided by the insight that full-rank structures that maximize parameters per unit of compute perform the best, we propose BTT-MoE, a novel Mixture-of-Experts (MoE) architecture obtained by sparsifying computation in the BTT structure. In contrast to the standard sparse MoE for each entire feed-forward network, BTT-MoE learns an MoE in every single linear layer of the model, including the projection matrices in the attention blocks. We find BTT-MoE provides a substantial compute-efficiency gain over dense layers and standard MoE.

comment: NeurIPS 2024. Code available at https://github.com/AndPotap/einsum-search

♻ ☆ Scattering Spectra Models for Physics

Sihao Cheng, Rudy Morel, Erwan Allys, Brice Ménard, Stéphane Mallat

comment: 11 pages, 6 figures, plus appendices, updated to published version

♻ ☆ Training Language Models to Self-Correct via Reinforcement Learning

Aviral Kumar, Vincent Zhuang, Rishabh Agarwal, Yi Su, John D Co-Reyes, Avi Singh, Kate Baumli, Shariq Iqbal, Colton Bishop, Rebecca Roelofs, Lei M Zhang, Kay McKinney, Disha Shrivastava, Cosmin Paduraru, George Tucker, Doina Precup, Feryal Behbahani, Aleksandra Faust

Self-correction is a highly desirable capability of large language models (LLMs), yet it has consistently been found to be largely ineffective in modern LLMs. Current methods for training self-correction typically depend on either multiple models, a more advanced model, or additional forms of supervision. To address these shortcomings, we develop a multi-turn online reinforcement learning (RL) approach, SCoRe, that significantly improves an LLM's self-correction ability using entirely self-generated data. To build SCoRe, we first show that variants of supervised fine-tuning (SFT) on offline model-generated correction traces are often insufficient for instilling self-correction behavior. In particular, we observe that training via SFT falls prey to either a distribution mismatch between mistakes made by the data-collection policy and the model's own responses, or to behavior collapse, where learning implicitly prefers only a certain mode of correction behavior that is often not effective at self-correction on test problems. SCoRe addresses these challenges by training under the model's own distribution of self-generated correction traces and using appropriate regularization to steer the learning process into learning a self-correction behavior that is effective at test time as opposed to fitting high-reward responses for a given prompt. This regularization process includes an initial phase of multi-turn RL on a base model to generate a policy initialization that is less susceptible to collapse, followed by using a reward bonus to amplify self-correction. With Gemini 1.0 Pro and 1.5 Flash models, we find that SCoRe achieves state-of-the-art self-correction performance, improving the base models' self-correction by 15.6% and 9.1% respectively on MATH and HumanEval.

♻ ☆ SaySelf: Teaching LLMs to Express Confidence with Self-Reflective Rationales EMNLP 2024

Tianyang Xu, Shujin Wu, Shizhe Diao, Xiaoze Liu, Xingyao Wang, Yangyi Chen, Jing Gao

comment: EMNLP 2024 Main

♻ ☆ Wrapper Boxes: Faithful Attribution of Model Predictions to Training Data

Yiheng Su, Junyi Jessy Li, Matthew Lease

Can we preserve the accuracy of neural models while also providing faithful explanations of model decisions to training data? We propose a "wrapper box'' pipeline: training a neural model as usual and then using its learned feature representation in classic, interpretable models to perform prediction. Across seven language models of varying sizes, including four large language models (LLMs), two datasets at different scales, three classic models, and four evaluation metrics, we first show that the predictive performance of wrapper classic models is largely comparable to the original neural models. Because classic models are transparent, each model decision is determined by a known set of training examples that can be directly shown to users. Our pipeline thus preserves the predictive performance of neural language models while faithfully attributing classic model decisions to training data. Among other use cases, such attribution enables model decisions to be contested based on responsible training instances. Compared to prior work, our approach achieves higher coverage and correctness in identifying which training data to remove to change a model decision. To reproduce findings, our source code is online at: https://github.com/SamSoup/WrapperBox.

♻ ☆ LInK: Learning Joint Representations of Design and Performance Spaces through Contrastive Learning for Mechanism Synthesis

Amin Heyrani Nobari, Akash Srivastava, Dan Gutfreund, Kai Xu, Faez Ahmed

In this paper, we introduce LInK, a novel framework that integrates contrastive learning of performance and design space with optimization techniques for solving complex inverse problems in engineering design with discrete and continuous variables. We focus on the path synthesis problem for planar linkage mechanisms. By leveraging a multimodal and transformation-invariant contrastive learning framework, LInK learns a joint representation that captures complex physics and design representations of mechanisms, enabling rapid retrieval from a vast dataset of over 10 million mechanisms. This approach improves precision through the warm start of a hierarchical unconstrained nonlinear optimization algorithm, combining the robustness of traditional optimization with the speed and adaptability of modern deep learning methods. Our results on an existing benchmark demonstrate that LInK outperforms existing methods with 28 times less error compared to a state of the art approach while taking 20 times less time on an existing benchmark. Moreover, we introduce a significantly more challenging benchmark, named LINK ABC, which involves synthesizing linkages that trace the trajectories of English capital alphabets, an inverse design benchmark task that existing methods struggle with due to large nonlinearities and tiny feasible space. Our results demonstrate that LInK not only advances the field of mechanism design but also broadens the applicability of contrastive learning and optimization to other areas of engineering. The code and data are publicly available at https://github.com/ahnobari/LInK.

♻ ☆ Language Model Empowered Spatio-Temporal Forecasting via Physics-Aware Reprogramming

Hao Wang, Jindong Han, Wei Fan, Hao Liu

♻ ☆ MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making

Yubin Kim, Chanwoo Park, Hyewon Jeong, Yik Siu Chan, Xuhai Xu, Daniel McDuff, Hyeonhoon Lee, Marzyeh Ghassemi, Cynthia Breazeal, Hae Won Park

♻ ☆ Unraveling the Truth: Do VLMs really Understand Charts? A Deep Dive into Consistency and Robustness

Srija Mukhopadhyay, Adnan Qidwai, Aparna Garimella, Pritika Ramu, Vivek Gupta, Dan Roth

comment: 22 pages, 9 Tables, 5 figures, 22 examples

♻ ☆ Navigable Graphs for High-Dimensional Nearest Neighbor Search: Constructions and Limits

Haya Diwan, Jinrui Gou, Cameron Musco, Christopher Musco, Torsten Suel

There has been significant recent interest in graph-based nearest neighbor search methods, many of which are centered on the construction of navigable graphs over high-dimensional point sets. A graph is navigable if we can successfully move from any starting node to any target node using a greedy routing strategy where we always move to the neighbor that is closest to the destination according to a given distance function. The complete graph is navigable for any point set, but the important question for applications is if sparser graphs can be constructed. While this question is fairly well understood in low-dimensions, we establish some of the first upper and lower bounds for high-dimensional point sets. First, we give a simple and efficient way to construct a navigable graph with average degree $O(\sqrt{n \log n })$ for any set of $n$ points, in any dimension, for any distance function. We compliment this result with a nearly matching lower bound: even under the Euclidean metric in $O(\log n)$ dimensions, a random point set has no navigable graph with average degree $O(n^{\alpha})$ for any $\alpha < 1/2$. Our lower bound relies on sharp anti-concentration bounds for binomial random variables, which we use to show that the near-neighborhoods of a set of random points do not overlap significantly, forcing any navigable graph to have many edges.

♻ ☆ Lessons Learned from a Unifying Empirical Study of Parameter-Efficient Transfer Learning (PETL) in Visual Recognition

Zheda Mai, Ping Zhang, Cheng-Hao Tu, Hong-You Chen, Li Zhang, Wei-Lun Chao

comment: Code is available at https://github.com/OSU-MLB/PETL_Vision

♻ ☆ Semi-Supervised Manifold Learning with Complexity Decoupled Chart Autoencoders

Stefan C. Schonsheck, Scott Mahan, Timo Klock, Alexander Cloninger, Rongjie Lai

Autoencoding is a popular method in representation learning. Conventional autoencoders employ symmetric encoding-decoding procedures and a simple Euclidean latent space to detect hidden low-dimensional structures in an unsupervised way. Some modern approaches to novel data generation such as generative adversarial networks askew this symmetry, but still employ a pair of massive networks--one to generate the image and another to judge the images quality based on priors learned from a training set. This work introduces a chart autoencoder with an asymmetric encoding-decoding process that can incorporate additional semi-supervised information such as class labels. Besides enhancing the capability for handling data with complicated topological and geometric structures, the proposed model can successfully differentiate nearby but disjoint manifolds and intersecting manifolds with only a small amount of supervision. Moreover, this model only requires a low-complexity encoding operation, such as a locally defined linear projection. We discuss the approximation power of such networks and derive a bound that essentially depends on the intrinsic dimension of the data manifold rather than the dimension of ambient space. Next we incorporate bounds for the sampling rate of training data need to faithfully represent a given data manifold. We present numerical experiments that verify that the proposed model can effectively manage data with multi-class nearby but disjoint manifolds of different classes, overlapping manifolds, and manifolds with non-trivial topology. Finally, we conclude with some experiments on computer vision and molecular dynamics problems which showcase the efficacy of our methods on real-world data.

♻ ☆ Probabilities of Chat LLMs Are Miscalibrated but Still Predict Correctness on Multiple-Choice Q&A

Benjamin Plaut, Nguyen X. Khanh, Tu Trinh

♻ ☆ Universal Functional Regression with Neural Operator Flows

Yaozhong Shi, Angela F. Gao, Zachary E. Ross, Kamyar Azizzadenesheli

Regression on function spaces is typically limited to models with Gaussian process priors. We introduce the notion of universal functional regression, in which we aim to learn a prior distribution over non-Gaussian function spaces that remains mathematically tractable for functional regression. To do this, we develop Neural Operator Flows (OpFlow), an infinite-dimensional extension of normalizing flows. OpFlow is an invertible operator that maps the (potentially unknown) data function space into a Gaussian process, allowing for exact likelihood estimation of functional point evaluations. OpFlow enables robust and accurate uncertainty quantification via drawing posterior samples of the Gaussian process and subsequently mapping them into the data function space. We empirically study the performance of OpFlow on regression and generation tasks with data generated from Gaussian processes with known posterior forms and non-Gaussian processes, as well as real-world earthquake seismograms with an unknown closed-form distribution.

♻ ☆ Identifying Factual Inconsistencies in Summaries: Grounding LLM Inference via Task Taxonomy EMNLP 2024

Liyan Xu, Zhenlin Su, Mo Yu, Jin Xu, Jinho D. Choi, Jie Zhou, Fei Liu

comment: Accepted to EMNLP 2024 Findings

♻ ☆ Explainable Earth Surface Forecasting under Extreme Events

Oscar J. Pellicer-Valero, Miguel-Ángel Fernández-Torres, Chaonan Ji, Miguel D. Mahecha, Gustau Camps-Valls

With climate change-related extreme events on the rise, high dimensional Earth observation data presents a unique opportunity for forecasting and understanding impacts on ecosystems. This is, however, impeded by the complexity of processing, visualizing, modeling, and explaining this data. To showcase how this challenge can be met, here we train a convolutional long short-term memory-based architecture on the novel DeepExtremeCubes dataset. DeepExtremeCubes includes around 40,000 long-term Sentinel-2 minicubes (January 2016-October 2022) worldwide, along with labeled extreme events, meteorological data, vegetation land cover, and topography map, sampled from locations affected by extreme climate events and surrounding areas. When predicting future reflectances and vegetation impacts through kernel normalized difference vegetation index, the model achieved an R$^2$ score of 0.9055 in the test set. Explainable artificial intelligence was used to analyze the model's predictions during the October 2020 Central South America compound heatwave and drought event. We chose the same area exactly one year before the event as counterfactual, finding that the average temperature and surface pressure are generally the best predictors under normal conditions. In contrast, minimum anomalies of evaporation and surface latent heat flux take the lead during the event. A change of regime is also observed in the attributions before the event, which might help assess how long the event was brewing before happening. The code to replicate all experiments and figures in this paper is publicly available at https://github.com/DeepExtremes/txyXAI

♻ ☆ Branches: A Fast Dynamic Programming and Branch & Bound Algorithm for Optimal Decision Trees

Ayman Chaouki, Jesse Read, Albert Bifet

Decision Tree (DT) Learning is a fundamental problem in Interpretable Machine Learning, yet it poses a formidable optimisation challenge. Despite numerous efforts dating back to the early 1990's, practical algorithms have only recently emerged, primarily leveraging Dynamic Programming (DP) and Branch & Bound (B&B) techniques. These methods fall into two categories: algorithms like DL8.5, MurTree and STreeD utilise an efficient DP strategy but lack effective bounds for pruning the search space; while algorithms like OSDT and GOSDT employ more efficient pruning bounds but at the expense of a less refined DP strategy. We introduce Branches, a new algorithm that combines the strengths of both approaches. Using DP and B&B with a novel analytical bound for efficient pruning, Branches offers both speed and sparsity optimisation. Unlike other methods, it also handles non-binary features. Theoretical analysis shows its lower complexity compared to existing methods, and empirical results confirm that Branches outperforms the state-of-the-art in speed, iterations, and optimality.

comment: This preprint is currently under review

♻ ☆ Graph Cuts with Arbitrary Size Constraints Through Optimal Transport

Chakib Fettal, Lazhar Labiod, Mohamed Nadif

A common way of partitioning graphs is through minimum cuts. One drawback of classical minimum cut methods is that they tend to produce small groups, which is why more balanced variants such as normalized and ratio cuts have seen more success. However, we believe that with these variants, the balance constraints can be too restrictive for some applications like for clustering of imbalanced datasets, while not being restrictive enough for when searching for perfectly balanced partitions. Here, we propose a new graph cut algorithm for partitioning graphs under arbitrary size constraints. We formulate the graph cut problem as a Gromov-Wasserstein with a concave regularizer problem. We then propose to solve it using an accelerated proximal GD algorithm which guarantees global convergence to a critical point, results in sparse solutions and only incurs an additional ratio of $\mathcal{O}(\log(n))$ compared to the classical spectral clustering algorithm but was seen to be more efficient.

comment: Published in Transactions on Machine Learning Research

♻ ☆ A Survey on Time-Series Pre-Trained Models

Qianli Ma, Zhen Liu, Zhenjing Zheng, Ziyang Huang, Siying Zhu, Zhongzhong Yu, James T. Kwok

Time-Series Mining (TSM) is an important research area since it shows great potential in practical applications. Deep learning models that rely on massive labeled data have been utilized for TSM successfully. However, constructing a large-scale well-labeled dataset is difficult due to data annotation costs. Recently, pre-trained models have gradually attracted attention in the time series domain due to their remarkable performance in computer vision and natural language processing. In this survey, we provide a comprehensive review of Time-Series Pre-Trained Models (TS-PTMs), aiming to guide the understanding, applying, and studying TS-PTMs. Specifically, we first briefly introduce the typical deep learning models employed in TSM. Then, we give an overview of TS-PTMs according to the pre-training techniques. The main categories we explore include supervised, unsupervised, and self-supervised TS-PTMs. Further, extensive experiments involving 27 methods, 434 datasets, and 679 transfer learning scenarios are conducted to analyze the advantages and disadvantages of transfer learning strategies, Transformer-based models, and representative TS-PTMs. Finally, we point out some potential directions of TS-PTMs for future work.

comment: Accepted in the IEEE Transactions on Knowledge and Data Engineering (TKDE)

♻ ☆ A Unified Theory of Quantum Neural Network Loss Landscapes

Eric R. Anschuetz

Classical neural networks with random initialization famously behave as Gaussian processes in the limit of many neurons, which allows one to completely characterize their training and generalization behavior. No such general understanding exists for quantum neural networks (QNNs), which -- outside of certain special cases -- are known to not behave as Gaussian processes when randomly initialized. We here prove that QNNs and their first two derivatives instead generally form what we call "Wishart processes," where certain algebraic properties of the network determine the hyperparameters of the process. This Wishart process description allows us to, for the first time: give necessary and sufficient conditions for a QNN architecture to have a Gaussian process limit; calculate the full gradient distribution, generalizing previously known barren plateau results; and calculate the local minima distribution of algebraically constrained QNNs. Our unified framework suggests a certain simple operational definition for the "trainability" of a given QNN model using a newly introduced, experimentally accessible quantity we call the "degrees of freedom" of the network architecture.

comment: 57 pages, 4 figures, added references and fixed minor bugs

♻ ☆ Linear combinations of Gaussian latents in generative models: interpolation and beyond

Erik Bodin, Carl Henrik Ek, Henry Moss

Sampling from generative models has become a crucial tool for applications like data synthesis and augmentation. Diffusion, Flow Matching and Continuous Normalizing Flows have shown effectiveness across various modalities, and rely on Gaussian latent variables for generation. For search-based or creative applications that require additional control over the generation process, it has become common to manipulate the latent variable directly. However, existing approaches for performing such manipulations (e.g. interpolation or forming low-dimensional representations) only work well in special cases or are network or data-modality specific. We propose Combination of Gaussian variables (COG) as a general purpose interpolation method that is easy to implement yet outperforms recent sophisticated methods. Moreover, COG naturally addresses the broader task of forming general linear combinations of latent variables, allowing the construction of subspaces of the latent space, dramatically simplifying the creation of expressive low-dimensional spaces of high-dimensional objects.

♻ ☆ Infinite Limits of Multi-head Transformer Dynamics

Blake Bordelon, Hamza Tahir Chaudhry, Cengiz Pehlevan

In this work, we analyze various scaling limits of the training dynamics of transformer models in the feature learning regime. We identify the set of parameterizations that admit well-defined infinite width and depth limits, allowing the attention layers to update throughout training--a relevant notion of feature learning in these models. We then use tools from dynamical mean field theory (DMFT) to analyze various infinite limits (infinite key/query dimension, infinite heads, and infinite depth) which have different statistical descriptions depending on which infinite limit is taken and how attention layers are scaled. We provide numerical evidence of convergence to the limits and discuss how the parameterization qualitatively influences learned features.

comment: Updating for Neurips 2024

♻ ☆ Towards Efficient Hyperdimensional Computing Using Photonics

Farbin Fayza, Cansu Demirkiran, Hanning Chen, Che-Kai Liu, Avi Mohan, Hamza Errahmouni, Sanggeon Yun, Mohsen Imani, David Zhang, Darius Bunandar, Ajay Joshi

♻ ☆ Avoiding Catastrophe in Online Learning by Asking for Help

Benjamin Plaut, Hanlin Zhu, Stuart Russell

Most learning algorithms with formal regret guarantees assume that no mistake is irreparable and essentially rely on trying all possible behaviors. This approach is problematic when some mistakes are \emph{catastrophic}, i.e., irreparable. We propose an online learning problem where the goal is to minimize the chance of catastrophe. Specifically, we assume that the payoff in each round represents the chance of avoiding catastrophe that round and aim to maximize the product of payoffs (the overall chance of avoiding catastrophe) while allowing a limited number of queries to a mentor. We first show that in general, any algorithm either constantly queries the mentor or is nearly guaranteed to cause catastrophe. However, in settings where the mentor policy class is learnable in the standard online learning model, we provide an algorithm whose regret and rate of querying the mentor both approach 0 as the time horizon grows. Conceptually, if a policy class is learnable in the absence of catastrophic risk, it is learnable in the presence of catastrophic risk if the agent can ask for help.

♻ ☆ Jailbreaking as a Reward Misspecification Problem

Zhihui Xie, Jiahui Gao, Lei Li, Zhenguo Li, Qi Liu, Lingpeng Kong

♻ ☆ Learning to Discretize Denoising Diffusion ODEs

Vinh Tong, Trung-Dung Hoang, Anji Liu, Guy Van den Broeck, Mathias Niepert

Diffusion Probabilistic Models (DPMs) are generative models showing competitive performance in various domains, including image synthesis and 3D point cloud generation. Sampling from pre-trained DPMs involves multiple neural function evaluations (NFE) to transform Gaussian noise samples into images, resulting in higher computational costs compared to single-step generative models such as GANs or VAEs. Therefore, reducing the number of NFEs while preserving generation quality is crucial. To address this, we propose LD3, a lightweight framework designed to learn the optimal time discretization for sampling. LD3 can be combined with various samplers and consistently improves generation quality without having to retrain resource-intensive neural networks. We demonstrate analytically and empirically that LD3 improves sampling efficiency with much less computational overhead. We evaluate our method with extensive experiments on 7 pre-trained models, covering unconditional and conditional sampling in both pixel-space and latent-space DPMs. We achieve FIDs of 2.38 (10 NFE), and 2.27 (10 NFE) on unconditional CIFAR10 and AFHQv2 in 5-10 minutes of training. LD3 offers an efficient approach to sampling from pre-trained diffusion models. Code is available at https://github.com/vinhsuhi/LD3/tree/main.

♻ ☆ Grounding Continuous Representations in Geometry: Equivariant Neural Fields

David R Wessels, David M Knigge, Samuele Papa, Riccardo Valperga, Sharvaree Vadgama, Efstratios Gavves, Erik J Bekkers

♻ ☆ One Policy to Run Them All: an End-to-end Learning Approach to Multi-Embodiment Locomotion

Nico Bohlinger, Grzegorz Czechmanowski, Maciej Krupka, Piotr Kicki, Krzysztof Walas, Jan Peters, Davide Tateo

Deep Reinforcement Learning techniques are achieving state-of-the-art results in robust legged locomotion. While there exists a wide variety of legged platforms such as quadruped, humanoids, and hexapods, the field is still missing a single learning framework that can control all these different embodiments easily and effectively and possibly transfer, zero or few-shot, to unseen robot embodiments. We introduce URMA, the Unified Robot Morphology Architecture, to close this gap. Our framework brings the end-to-end Multi-Task Reinforcement Learning approach to the realm of legged robots, enabling the learned policy to control any type of robot morphology. The key idea of our method is to allow the network to learn an abstract locomotion controller that can be seamlessly shared between embodiments thanks to our morphology-agnostic encoders and decoders. This flexible architecture can be seen as a potential first step in building a foundation model for legged robot locomotion. Our experiments show that URMA can learn a locomotion policy on multiple embodiments that can be easily transferred to unseen robot platforms in simulation and the real world.

♻ ☆ To Know or Not To Know? Analyzing Self-Consistency of Large Language Models under Ambiguity EMNLP 2024

Anastasiia Sedova, Robert Litschko, Diego Frassinelli, Benjamin Roth, Barbara Plank

comment: EMNLP 2024 Findings

♻ ☆ FusionDTI: Fine-grained Binding Discovery with Token-level Fusion for Drug-Target Interaction

Zhaohan Meng, Zaiqiao Meng, Ke Yuan, Iadh Ounis

Predicting drug-target interaction (DTI) is critical in the drug discovery process. Despite remarkable advances in recent DTI models through the integration of representations from diverse drug and target encoders, such models often struggle to capture the fine-grained interactions between drugs and protein, i.e. the binding of specific drug atoms (or substructures) and key amino acids of proteins, which is crucial for understanding the binding mechanisms and optimising drug design. To address this issue, this paper introduces a novel model, called FusionDTI, which uses a token-level Fusion module to effectively learn fine-grained information for Drug-Target Interaction. In particular, our FusionDTI model uses the SELFIES representation of drugs to mitigate sequence fragment invalidation and incorporates the structure-aware (SA) vocabulary of target proteins to address the limitation of amino acid sequences in structural information, additionally leveraging pre-trained language models extensively trained on large-scale biomedical datasets as encoders to capture the complex information of drugs and targets. Experiments on three well-known benchmark datasets show that our proposed FusionDTI model achieves the best performance in DTI prediction compared with seven existing state-of-the-art baselines. Furthermore, our case study indicates that FusionDTI could highlight the potential binding sites, enhancing the explainability of the DTI prediction.

comment: 10 pages, 8 figures

♻ ☆ A large-scale operational study of fingerprint quality and demographics

Javier Galbally, Aleksandrs Cepilovs, Ramon Blanco-Gonzalo, Gillian Ormiston, Oscar Miguel-Hurtado, Istvan Sz. Racz

♻ ☆ Recurrent Interpolants for Probabilistic Time Series Prediction

Yu Chen, Marin Biloš, Sarthak Mittal, Wei Deng, Kashif Rasul, Anderson Schneider

Sequential models like recurrent neural networks and transformers have become standard for probabilistic multivariate time series forecasting across various domains. Despite their strengths, they struggle with capturing high-dimensional distributions and cross-feature dependencies. Recent work explores generative approaches using diffusion or flow-based models, extending to time series imputation and forecasting. However, scalability remains a challenge. This work proposes a novel method combining recurrent neural networks' efficiency with diffusion models' probabilistic modeling, based on stochastic interpolants and conditional generation with control features, offering insights for future developments in this dynamic field.

♻ ☆ Consensus Knowledge Graph Learning via Multi-view Sparse Low Rank Block Model

Tianxi Cai, Dong Xia, Luwan Zhang, Doudou Zhou

Network analysis has been a powerful tool to unveil relationships and interactions among a large number of objects. Yet its effectiveness in accurately identifying important node-node interactions is challenged by the rapidly growing network size, with data being collected at an unprecedented granularity and scale. Common wisdom to overcome such high dimensionality is collapsing nodes into smaller groups and conducting connectivity analysis on the group level. Dividing efforts into two phases inevitably opens a gap in consistency and drives down efficiency. Consensus learning emerges as a new normal for common knowledge discovery with multiple data sources available. In this paper, we propose a unified multi-view sparse low-rank block model (msLBM) framework, which enables simultaneous grouping and connectivity analysis by combining multiple data sources. The msLBM framework efficiently represents overlapping information across large scale concepts and accommodates different types of heterogeneity across sources. Both features are desirable when analyzing high dimensional electronic health record (EHR) datasets from multiple health systems. An estimating procedure based on the alternating minimization algorithm is proposed. Our theoretical results demonstrate that a consensus knowledge graph can be more accurately learned by leveraging multi-source datasets, and statistically optimal rates can be achieved under mild conditions. Applications to the real world EHR data suggest that our proposed msLBM algorithm can more reliably reveal network structure among clinical concepts by effectively combining summary level EHR data from multiple health systems.

♻ ☆ Epistemic Monte Carlo Tree Search

Yaniv Oren, Villiam Vadocz, Matthijs T. J. Spaan, Wendelin Böhmer

The AlphaZero/MuZero (A/MZ) family of algorithms has achieved remarkable success across various challenging domains by integrating Monte Carlo Tree Search (MCTS) with learned models. Learned models introduce epistemic uncertainty, which is caused by learning from limited data and is useful for exploration in sparse reward environments. MCTS does not account for the propagation of this uncertainty however. To address this, we introduce Epistemic MCTS (EMCTS): a theoretically motivated approach to account for the epistemic uncertainty in search and harness the search for deep exploration. In the challenging sparse-reward task of writing code in the Assembly language SUBLEQ, AZ paired with our method achieves significantly higher sample efficiency over baseline AZ. Search with EMCTS solves variations of the commonly used hard-exploration benchmark Deep Sea - which baseline A/MZ are practically unable to solve - much faster than an otherwise equivalent method that does not use search for uncertainty estimation, demonstrating significant benefits from search for epistemic uncertainty estimation.

♻ ☆ Any-Quantile Probabilistic Forecasting of Short-Term Electricity Demand

Slawek Smyl, Boris N. Oreshkin, Paweł Pełka, Grzegorz Dudek

Power systems operate under uncertainty originating from multiple factors that are impossible to account for deterministically. Distributional forecasting is used to control and mitigate risks associated with this uncertainty. Recent progress in deep learning has helped to significantly improve the accuracy of point forecasts, while accurate distributional forecasting still presents a significant challenge. In this paper, we propose a novel general approach for distributional forecasting capable of predicting arbitrary quantiles. We show that our general approach can be seamlessly applied to two distinct neural architectures leading to the state-of-the-art distributional forecasting results in the context of short-term electricity demand forecasting task. We empirically validate our method on 35 hourly electricity demand time-series for European countries. Our code is available here: https://github.com/boreshkinai/any-quantile.

♻ ☆ Rapid and Precise Topological Comparison with Merge Tree Neural Networks IEEE VIS 2024

Yu Qin, Brittany Terese Fasy, Carola Wenk, Brian Summa

Merge trees are a valuable tool in the scientific visualization of scalar fields; however, current methods for merge tree comparisons are computationally expensive, primarily due to the exhaustive matching between tree nodes. To address this challenge, we introduce the Merge Tree Neural Network (MTNN), a learned neural network model designed for merge tree comparison. The MTNN enables rapid and high-quality similarity computation. We first demonstrate how to train graph neural networks, which emerged as effective encoders for graphs, in order to produce embeddings of merge trees in vector spaces for efficient similarity comparison. Next, we formulate the novel MTNN model that further improves the similarity comparisons by integrating the tree and node embeddings with a new topological attention mechanism. We demonstrate the effectiveness of our model on real-world data in different domains and examine our model's generalizability across various datasets. Our experimental analysis demonstrates our approach's superiority in accuracy and efficiency. In particular, we speed up the prior state-of-the-art by more than $100\times$ on the benchmark datasets while maintaining an error rate below $0.1\%$.

comment: Published on IEEE VIS 2024 with Best Paper Award

♻ ☆ Editable Concept Bottleneck Models

Lijie Hu, Chenyang Ren, Zhengyu Hu, Hongbin Lin, Cheng-Long Wang, Hui Xiong, Jingfeng Zhang, Di Wang

comment: 36 pages

♻ ☆ Truncated Kernel Stochastic Gradient Descent on Spheres

JinHui Bai, Lei Shi

Inspired by the structure of spherical harmonics, we propose the truncated kernel stochastic gradient descent (T-kernel SGD) algorithm with a least-square loss function for spherical data fitting. T-kernel SGD employs a "truncation" operation, enabling the application of series-based kernels function in stochastic gradient descent, thereby avoiding the difficulties of finding suitable closed-form kernel functions in high-dimensional spaces. In contrast to traditional kernel SGD, T-kernel SGD is more effective in balancing bias and variance by dynamically adjusting the hypothesis space during iterations. The most significant advantage of the proposed algorithm is that it can achieve theoretically optimal convergence rates using a constant step size (independent of the sample size) while overcoming the inherent saturation problem of kernel SGD. Additionally, we leverage the structure of spherical polynomials to derive an equivalent T-kernel SGD, significantly reducing storage and computational costs compared to kernel SGD. Typically, T-kernel SGD requires only $\mathcal{O}(n^{1+\frac{d}{d-1}\epsilon})$ computational complexity and $\mathcal{O}(n^{\frac{d}{d-1}\epsilon})$ storage to achieve optimal rates for the d-dimensional sphere, where $0<\epsilon<\frac{1}{2}$ can be arbitrarily small if the optimal fitting or the underlying space possesses sufficient regularity. This regularity is determined by the smoothness parameter of the objective function and the decaying rate of the eigenvalues of the integral operator associated with the kernel function, both of which reflect the difficulty of the estimation problem. Our main results quantitatively characterize how this prior information influences the convergence of T-kernel SGD. The numerical experiments further validate the theoretical findings presented in this paper.

comment: 57 pages, 7 figures

♻ ☆ Densely Multiplied Physics Informed Neural Networks

Feilong Jiang, Xiaonan Hou, Min Xia

Although physics-informed neural networks (PINNs) have shown great potential in dealing with nonlinear partial differential equations (PDEs), it is common that PINNs will suffer from the problem of insufficient precision or obtaining incorrect outcomes. Unlike most of the existing solutions trying to enhance the ability of PINN by optimizing the training process, this paper improved the neural network architecture to improve the performance of PINN. We propose a densely multiply PINN (DM-PINN) architecture, which multiplies the output of a hidden layer with the outputs of all the behind hidden layers. Without introducing more trainable parameters, this effective mechanism can significantly improve the accuracy of PINNs. The proposed architecture is evaluated on four benchmark examples (Allan-Cahn equation, Helmholtz equation, Burgers equation and 1D convection equation). Comparisons between the proposed architecture and different PINN structures demonstrate the superior performance of the DM-PINN in both accuracy and efficiency.

comment: 15 pages, 9 figures

♻ ☆ Jacobian Descent for Multi-Objective Optimization

Pierre Quinton, Valérian Rey

Many optimization problems require balancing multiple conflicting objectives. As gradient descent is limited to single-objective optimization, we introduce its direct generalization: Jacobian descent (JD). This algorithm iteratively updates parameters using the Jacobian matrix of a vector-valued objective function, in which each row is the gradient of an individual objective. While several methods to combine gradients already exist in the literature, they are generally hindered when the objectives conflict. In contrast, we propose projecting gradients to fully resolve conflict while ensuring that they preserve an influence proportional to their norm. We prove significantly stronger convergence guarantees with this approach, supported by our empirical results. Our method also enables instance-wise risk minimization (IWRM), a novel learning paradigm in which the loss of each training example is considered a separate objective. Applied to simple image classification tasks, IWRM exhibits promising results compared to the direct minimization of the average loss. Additionally, we outline an efficient implementation of JD using the Gramian of the Jacobian matrix to reduce time and memory requirements.

comment: 39 pages, 10 figures, conference

♻ ☆ Logistic Regression makes small LLMs strong and explainable "tens-of-shot" classifiers

Marcus Buckmann, Edward Hill

comment: 48 pages, 24 figures

♻ ☆ Self-Training for Sample-Efficient Active Learning for Text Classification with Pre-Trained Language Models EMNLP 2024

Christopher Schröder, Gerhard Heyer

comment: Accepted to EMNLP 2024

♻ ☆ G-Transformer for Conditional Average Potential Outcome Estimation over Time

Konstantin Hess, Dennis Frauen, Valentyn Melnychuk, Stefan Feuerriegel

Estimating potential outcomes for treatments over time based on observational data is important for personalized decision-making in medicine. Yet, existing neural methods for this task either (1) do not perform proper adjustments for time-varying confounders, or (2) suffer from large estimation variance. In order to address both limitations, we introduce the G-transformer (GT). Our GT is a novel, neural end-to-end model which adjusts for time-varying confounders, and provides low-variance estimation of conditional average potential outcomes (CAPOs) over time. Specifically, our GT is the first neural model to perform regression-based iterative G-computation for CAPOs in the time-varying setting. We evaluate the effectiveness of our GT across various experiments. In sum, this work represents a significant step towards personalized decision-making from electronic health records.

♻ ☆ Which Experiences Are Influential for RL Agents? Efficiently Estimating The Influence of Experiences

Takuya Hiraoka, Guanquan Wang, Takashi Onishi, Yoshimasa Tsuruoka

In reinforcement learning (RL) with experience replay, experiences stored in a replay buffer influence the RL agent's performance. Information about how these experiences influence the agent's performance is valuable for various purposes, such as identifying experiences that negatively influence underperforming agents. One method for estimating the influence of experiences is the leave-one-out (LOO) method. However, this method is usually computationally prohibitive. In this paper, we present Policy Iteration with Turn-over Dropout (PIToD), which efficiently estimates the influence of experiences. We evaluate how accurately PIToD estimates the influence of experiences and its efficiency compared to LOO. We then apply PIToD to amend underperforming RL agents, i.e., we use PIToD to estimate negatively influential experiences for the RL agents and to delete the influence of these experiences. We show that RL agents' performance is significantly improved via amendments with PIToD.

comment: Source code: https://github.com/TakuyaHiraoka/Which-Experiences-Are-Influential-for-RL-Agents

♻ ☆ AQMLator -- An Auto Quantum Machine Learning E-Platform

Tomasz Rybotycki, Piotr Gawron

comment: 15 pages, 3 figures, links to software in the text

♻ ☆ Is In-Context Learning Sufficient for Instruction Following in LLMs?

Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion

comment: Preprint. Code at https://github.com/tml-epfl/icl-alignment

♻ ☆ Comparison of Reservoir Computing topologies using the Recurrent Kernel approach

Giuseppe Alessio D'Inverno, Jonathan Dong

Reservoir Computing (RC) has become popular in recent years thanks to its fast and efficient computational capabilities. Standard RC has been shown to be equivalent in the asymptotic limit to Recurrent Kernels, which helps in analyzing its expressive power. However, many well-established RC paradigms, such as Leaky RC, Sparse RC, and Deep RC, are yet to be systematically analyzed in such a way. We define the Recurrent Kernel limit of all these RC topologies and conduct a convergence study for a wide range of activation functions and hyperparameters. Our findings provide new insights into various aspects of Reservoir Computing. First, we demonstrate that there is an optimal sparsity level which grows with the reservoir size. Furthermore, our analysis suggests that Deep RC should use reservoir layers of decreasing sizes. Finally, we perform a benchmark demonstrating the efficiency of Structured Reservoir Computing compared to vanilla and Sparse Reservoir Computing.

♻ ☆ Probabilistic forecasting of power system imbalance using neural network-based ensembles

Jonas Van Gompel, Bert Claessens, Chris Develder

Keeping the balance between electricity generation and consumption is becoming increasingly challenging and costly, mainly due to the rising share of renewables, electric vehicles and heat pumps and electrification of industrial processes. Accurate imbalance forecasts, along with reliable uncertainty estimations, enable transmission system operators (TSOs) to dispatch appropriate reserve volumes, reducing balancing costs. Further, market parties can use these probabilistic forecasts to design strategies that exploit asset flexibility to help balance the grid, generating revenue with known risks. Despite its importance, literature regarding system imbalance (SI) forecasting is limited. Further, existing methods do not focus on situations with high imbalance magnitude, which are crucial to forecast accurately for both TSOs and market parties. Hence, we propose an ensemble of C-VSNs, which are our adaptation of variable selection networks (VSNs). Each minute, our model predicts the imbalance of the current and upcoming two quarter-hours, along with uncertainty estimations on these forecasts. We evaluate our approach by forecasting the imbalance of Belgium, where high imbalance magnitude is defined as $|$SI$| > 500\,$MW (occurs 1.3% of the time in Belgium). For high imbalance magnitude situations, our model outperforms the state-of-the-art by 23.4% (in terms of continuous ranked probability score (CRPS), which evaluates probabilistic forecasts), while also attaining a 6.5% improvement in overall CRPS. Similar improvements are achieved in terms of root-mean-squared error. Additionally, we developed a fine-tuning methodology to effectively include new inputs with limited history in our model. This work was performed in collaboration with Elia (the Belgian TSO) to further improve their imbalance forecasts, demonstrating the relevance of our work.

♻ ☆ The Vital Role of Gradient Clipping in Byzantine-Resilient Distributed Learning

Youssef Allouah, Rachid Guerraoui, Nirupam Gupta, Ahmed Jellouli, Geovani Rizk, John Stephan

Byzantine-resilient distributed machine learning seeks to achieve robust learning performance in the presence of misbehaving or adversarial workers. While state-of-the-art (SOTA) robust distributed gradient descent (Robust-DGD) methods were proven theoretically optimal, their empirical success has often relied on pre-aggregation gradient clipping. However, the currently considered static clipping strategy exhibits mixed results: improving robustness against some attacks while being ineffective or detrimental against others. We address this gap by proposing a principled adaptive clipping strategy, termed Adaptive Robust Clipping (ARC). We show that ARC consistently enhances the empirical robustness of SOTA Robust-DGD methods, while preserving the theoretical robustness guarantees. Our analysis shows that ARC provably improves the asymptotic convergence guarantee of Robust-DGD in the case when the model is well-initialized. We validate this theoretical insight through an exhaustive set of experiments on benchmark image classification tasks. We observe that the improvement induced by ARC is more pronounced in highly heterogeneous and adversarial settings.

♻ ☆ Reinforcement learning for anisotropic p-adaptation and error estimation in high-order solvers

David Huergo, Martín de Frutos, Eduardo Jané, Oscar A. Marino, Gonzalo Rubio, Esteban Ferrer

We present a novel approach to automate and optimize anisotropic p-adaptation in high-order h/p solvers using Reinforcement Learning (RL). The dynamic RL adaptation uses the evolving solution to adjust the high-order polynomials. We develop an offline training approach, decoupled from the main solver, which shows minimal overcost when performing simulations. In addition, we derive an inexpensive RL-based error estimation approach that enables the quantification of local discretization errors. The proposed methodology is agnostic to both the computational mesh and the partial differential equation to be solved. The application of RL to mesh adaptation offers several benefits. It enables automated and adaptive mesh refinement, reducing the need for manual intervention. It optimizes computational resources by dynamically allocating high-order polynomials where necessary and minimizing refinement in stable regions. This leads to computational cost savings while maintaining the accuracy of the solution. Furthermore, RL allows for the exploration of unconventional mesh adaptations, potentially enhancing the accuracy and robustness of simulations. This work extends our original research, offering a more robust, reproducible, and generalizable approach applicable to complex three-dimensional problems. We provide validation for laminar and turbulent cases: circular cylinders, Taylor Green Vortex and a 10MW wind turbine to illustrate the flexibility of the proposed approach.

comment: 38 pages, 18 figures, 5 tables

♻ ☆ GROD: Enhancing Generalization of Transformer with Out-of-Distribution Detection

Yijin Zhou, Yuguang Wang

Transformer networks excel in natural language processing (NLP) and computer vision (CV) tasks. However, they face challenges in generalizing to Out-of-Distribution (OOD) datasets, that is, data whose distribution differs from that seen during training. The OOD detection aims to distinguish data that deviates from the expected distribution, while maintaining optimal performance on in-distribution (ID) data. This paper introduces a novel approach based on OOD detection, termed the Generate Rounded OOD Data (GROD) algorithm, which significantly bolsters the generalization performance of transformer networks across various tasks. GROD is motivated by our new OOD detection Probably Approximately Correct (PAC) Theory for transformer. The transformer has learnability in terms of OOD detection that is, when the data is sufficient the outlier can be well represented. By penalizing the misclassification of OOD data within the loss function and generating synthetic outliers, GROD guarantees learnability and refines the decision boundaries between inlier and outlier. This strategy demonstrates robust adaptability and general applicability across different data types. Evaluated across diverse OOD detection tasks in NLP and CV, GROD achieves SOTA regardless of data format. The code is available at https://anonymous.4open.science/r/GROD-OOD-Detection-with-transformers-B70F.

♻ ☆ Deep Model Interpretation with Limited Data : A Coreset-based Approach

Hamed Behzadi-Khormouji, José Oramas

♻ ☆ Efficient Finite Initialization for Tensorized Neural Networks

Alejandro Mata Ali, Iñigo Perez Delgado, Marina Ristol Roura, Aitor Moreno Fdez. de Leceta

We present a novel method for initializing layers of tensorized neural networks in a way that avoids the explosion of the parameters of the matrix it emulates. The method is intended for layers with a high number of nodes in which there is a connection to the input or output of all or most of the nodes, we cannot or do not want to store/calculate all the elements of the represented layer and they follow a smooth distribution. This method is equally applicable to normalize general tensor networks in which we want to avoid overflows. The core of this method is the use of the Frobenius norm and the partial lineal entrywise norm of reduced forms of the layer in an iterative partial form, so that it has to be finite and within a certain range. These norms are efficient to compute, fully or partially for most cases of interest. In addition, the method benefits from the reuse of intermediate calculations. We apply the method to different layers and check its performance. We create a Python function to run it on an arbitrary layer, available in a Jupyter Notebook in the i3BQuantum repository: https://github.com/i3BQuantumTeam/Q4Real/blob/e07c827651ef16bcf74590ab965ea3985143f891/Quantum-Inspired%20Variational%20Methods/TN_Normalizer.ipynb

comment: 8 pages, 13 figures

♻ ☆ Semi-Supervised Fine-Tuning of Vision Foundation Models with Content-Style Decomposition

Mariia Drozdova, Vitaliy Kinakh, Yury Belousov, Erica Lastufka, Slava Voloshynovskiy

comment: preprint

♻ ☆ Conjugate Bayesian Two-step Change Point Detection for Hawkes Process NeurIPS 2024

Zeyue Zhang, Xiaoling Lu, Feng Zhou

The Bayesian two-step change point detection method is popular for the Hawkes process due to its simplicity and intuitiveness. However, the non-conjugacy between the point process likelihood and the prior requires most existing Bayesian two-step change point detection methods to rely on non-conjugate inference methods. These methods lack analytical expressions, leading to low computational efficiency and impeding timely change point detection. To address this issue, this work employs data augmentation to propose a conjugate Bayesian two-step change point detection method for the Hawkes process, which proves to be more accurate and efficient. Extensive experiments on both synthetic and real data demonstrate the superior effectiveness and efficiency of our method compared to baseline methods. Additionally, we conduct ablation studies to explore the robustness of our method concerning various hyperparameters. Our code is publicly available at https://github.com/Aurora2050/CoBay-CPD.

comment: 10 pages, accepted by NeurIPS 2024

♻ ☆ Major Entity Identification: A Generalizable Alternative to Coreference Resolution

Kawshik Manikantan, Shubham Toshniwal, Makarand Tapaswi, Vineet Gandhi

comment: 17 pages, 6 figures

♻ ☆ C-MELT: Contrastive Enhanced Masked Auto-Encoders for ECG-Language Pre-Training

Manh Pham, Aaqib Saeed, Dong Ma

♻ ☆ Foundational Inference Models for Dynamical Systems

Patrick Seifner, Kostadin Cvejoski, Antonia Körner, Ramsés J. Sánchez

Dynamical systems governed by ordinary differential equations (ODEs) serve as models for a vast number of natural and social phenomena. In this work, we offer a fresh perspective on the classical problem of imputing missing time series data, whose underlying dynamics are assumed to be determined by ODEs. Specifically, we revisit ideas from amortized inference and neural operators, and propose a novel supervised learning framework for zero-shot time series imputation, through parametric functions satisfying some (hidden) ODEs. Our proposal consists of two components. First, a broad probability distribution over the space of ODE solutions, observation times and noise mechanisms, with which we generate a large, synthetic dataset of (hidden) ODE solutions, along with their noisy and sparse observations. Second, a neural recognition model that is trained offline, to map the generated time series onto the spaces of initial conditions and time derivatives of the (hidden) ODE solutions, which we then integrate to impute the missing data. We empirically demonstrate that one and the same (pretrained) recognition model can perform zero-shot imputation across 63 distinct time series with missing values, each sampled from widely different dynamical systems. Likewise, we demonstrate that it can perform zero-shot imputation of missing high-dimensional data in 10 vastly different settings, spanning human motion, air quality, traffic and electricity studies, as well as Navier-Stokes simulations -- without requiring any fine-tuning. What is more, our proposal often outperforms state-of-the-art methods, which are trained on the target datasets. Our pretrained model will be available online soon.

♻ ☆ Integer-only Quantized Transformers for Embedded FPGA-based Time-series Forecasting in AIoT

Tianheng Ling, Chao Qian, Gregor Schiele

This paper presents the design of a hardware accelerator for Transformers, optimized for on-device time-series forecasting in AIoT systems. It integrates integer-only quantization and Quantization-Aware Training with optimized hardware designs to realize 6-bit and 4-bit quantized Transformer models, which achieved precision comparable to 8-bit quantized models from related research. Utilizing a complete implementation on an embedded FPGA (Xilinx Spartan-7 XC7S15), we examine the feasibility of deploying Transformer models on embedded IoT devices. This includes a thorough analysis of achievable precision, resource utilization, timing, power, and energy consumption for on-device inference. Our results indicate that while sufficient performance can be attained, the optimization process is not trivial. For instance, reducing the quantization bitwidth does not consistently result in decreased latency or energy consumption, underscoring the necessity of systematically exploring various optimization combinations. Compared to an 8-bit quantized Transformer model in related studies, our 4-bit quantized Transformer model increases test loss by only 0.63%, operates up to 132.33x faster, and consumes 48.19x less energy.

comment: Accepted by 2024 IEEE Annual Congress on Artificial Intelligence of Things (IEEE AIoT) and got best paper award. 7 pages, 3 figures, 4 tables

♻ ☆ Informed deep hierarchical classification: a non-standard analysis inspired approach

Lorenzo Fiaschi, Marco Cococcioni

This work proposes a novel approach to the deep hierarchical classification task, i.e., the problem of classifying data according to multiple labels organized in a rigid parent-child structure. It consists in a multi-output deep neural network equipped with specific projection operators placed before each output layer. The design of such an architecture, called lexicographic hybrid deep neural network (LH-DNN), has been possible by combining tools from different and quite distant research fields: lexicographic multi-objective optimization, non-standard analysis, and deep learning. To assess the efficacy of the approach, the resulting network is compared against the B-CNN, a convolutional neural network tailored for hierarchical classification tasks, on the CIFAR10, CIFAR100 (where it has been originally and recently proposed before being adopted and tuned for multiple real-world applications) and Fashion-MNIST benchmarks. Evidence states that an LH-DNN can achieve comparable if not superior performance, especially in the learning of the hierarchical relations, in the face of a drastic reduction of the learning parameters, training epochs, and computational time, without the need for ad-hoc loss functions weighting values.

♻ ☆ Cometh: A continuous-time discrete-state graph diffusion model

Antoine Siraudin, Fragkiskos D. Malliaros, Christopher Morris

Discrete-state denoising diffusion models led to state-of-the-art performance in graph generation, especially in the molecular domain. Recently, they have been transposed to continuous time, allowing more flexibility in the reverse process and a better trade-off between sampling efficiency and quality. Here, to leverage the benefits of both approaches, we propose Cometh, a continuous-time discrete-state graph diffusion model, tailored to the specificities of graph data. In addition, we also successfully replaced the set of structural encodings previously used in the discrete graph diffusion model with a single random-walk-based encoding, providing a simple and principled way to boost the model's expressive power. Empirically, we show that integrating continuous time leads to significant improvements across various metrics over state-of-the-art discrete-state diffusion models on a large set of molecular and non-molecular benchmark datasets. In terms of VUN samples, Cometh obtains a near-perfect performance of 99.5% on the planar graph dataset and outperforms DiGress by 12.6% on the large GuacaMol dataset.

comment: 23 pages

♻ ☆ Robustness investigation of cross-validation based quality measures for model assessment

Thomas Most, Lars Gräning, Sebastian Wolff

In this paper the accuracy and robustness of quality measures for the assessment of machine learning models are investigated. The prediction quality of a machine learning model is evaluated model-independent based on a cross-validation approach, where the approximation error is estimated for unknown data. The presented measures quantify the amount of explained variation in the model prediction. The reliability of these measures is assessed by means of several numerical examples, where an additional data set for the verification of the estimated prediction error is available. Furthermore, the confidence bounds of the presented quality measures are estimated and local quality measures are derived from the prediction residuals obtained by the cross-validation approach.

comment: accepted for publication in Engineering Modelling, Analysis & Simulation (EMAS)

♻ ☆ Simple Drop-in LoRA Conditioning on Attention Layers Will Improve Your Diffusion Model

Joo Young Choi, Jaesung R. Park, Inkyu Park, Jaewoong Cho, Albert No, Ernest K. Ryu

♻ ★ HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models

Seanie Lee, Haebin Seong, Dong Bok Lee, Minki Kang, Xiaoyin Chen, Dominik Wagner, Yoshua Bengio, Juho Lee, Sung Ju Hwang

♻ ☆ GRVFL-MV: Graph Random Vector Functional Link Based on Multi-View Learning

M. Tanveer, R. K. Sharma, M. Sajid, A. Quadir

The classification performance of the random vector functional link (RVFL), a randomized neural network, has been widely acknowledged. However, due to its shallow learning nature, RVFL often fails to consider all the relevant information available in a dataset. Additionally, it overlooks the geometrical properties of the dataset. To address these limitations, a novel graph random vector functional link based on multi-view learning (GRVFL-MV) model is proposed. The proposed model is trained on multiple views, incorporating the concept of multiview learning (MVL), and it also incorporates the geometrical properties of all the views using the graph embedding (GE) framework. The fusion of RVFL networks, MVL, and GE framework enables our proposed model to achieve the following: i) efficient learning: by leveraging the topology of RVFL, our proposed model can efficiently capture nonlinear relationships within the multi-view data, facilitating efficient and accurate predictions; ii) comprehensive representation: fusing information from diverse perspectives enhance the proposed model's ability to capture complex patterns and relationships within the data, thereby improving the model's overall generalization performance; and iii) structural awareness: by employing the GE framework, our proposed model leverages the original data distribution of the dataset by naturally exploiting both intrinsic and penalty subspace learning criteria. The evaluation of the proposed GRVFL-MV model on various datasets, including 27 UCI and KEEL datasets, 50 datasets from Corel5k, and 45 datasets from AwA, demonstrates its superior performance compared to baseline models. These results highlight the enhanced generalization capabilities of the proposed GRVFL-MV model across a diverse range of datasets.

♻ ☆ What Can We Learn from State Space Models for Machine Learning on Graphs?

Yinan Huang, Siqi Miao, Pan Li

Machine learning on graphs has recently found extensive applications across domains. However, the commonly used Message Passing Neural Networks (MPNNs) suffer from limited expressive power and struggle to capture long-range dependencies. Graph transformers offer a strong alternative due to their global attention mechanism, but they come with great computational overheads, especially for large graphs. In recent years, State Space Models (SSMs) have emerged as a compelling approach to replace full attention in transformers to model sequential data. It blends the strengths of RNNs and CNNs, offering a) efficient computation, b) the ability to capture long-range dependencies, and c) good generalization across sequences of various lengths. However, extending SSMs to graph-structured data presents unique challenges due to the lack of canonical node ordering in graphs. In this work, we propose Graph State Space Convolution (GSSC) as a principled extension of SSMs to graph-structured data. By leveraging global permutation-equivariant set aggregation and factorizable graph kernels that rely on relative node distances as the convolution kernels, GSSC preserves all three advantages of SSMs. We demonstrate the provably stronger expressiveness of GSSC than MPNNs in counting graph substructures and show its effectiveness across 11 real-world, widely used benchmark datasets. GSSC achieves the best results on 6 out of 11 datasets with all significant improvements compared to the state-of-the-art baselines and second-best results on the other 5 datasets. Our findings highlight the potential of GSSC as a powerful and scalable model for graph machine learning. Our code is available at https://github.com/Graph-COM/GSSC.

♻ ☆ Identifying Three-Dimensional Radiative Patterns Associated with Early Tropical Cyclone Intensification

Frederick Iat-Hin Tam, Tom Beucler, James H. Ruppert Jr

Cloud radiative feedback impacts early tropical cyclone (TC) intensification, but limitations in existing diagnostic frameworks make them unsuitable for studying asymmetric or transient radiative heating. We propose a linear Variational Encoder-Decoder (VED) to learn the hidden relationship between radiation and the surface intensification of realistic simulated TCs. Limiting VED model inputs enables using its uncertainty to identify periods when radiation has more importance for intensification. A close examination of the extracted 3D radiative structures suggests that longwave radiative forcing from inner core deep convection and shallow clouds both contribute to intensification, with the deep convection having the most impact overall. We find that deep convection downwind of the shallow clouds is critical to the intensification of Haiyan. Our work demonstrates that machine learning can discover thermodynamic-kinematic relationships without relying on axisymmetric or deterministic assumptions, paving the way towards the objective discovery of processes leading to TC intensification in realistic conditions.

comment: 15 pages, 6 figures (main text)

Multimedia 9

☆ Does SpatioTemporal information benefit Two video summarization benchmarks? ECAI 2024

Aashutosh Ganesh, Mirela Popa, Daan Odijk, Nava Tintarev

comment: Accepted for presentation at AEQUITAS workshop, Co-located with ECAI 2024

☆ Enriching Music Descriptions with a Finetuned-LLM and Metadata for Text-to-Music Retrieval ICASSP 2024

SeungHeon Doh, Minhee Lee, Dasaem Jeong, Juhan Nam

Text-to-Music Retrieval, finding music based on a given natural language query, plays a pivotal role in content discovery within extensive music databases. To address this challenge, prior research has predominantly focused on a joint embedding of music audio and text, utilizing it to retrieve music tracks that exactly match descriptive queries related to musical attributes (i.e. genre, instrument) and contextual elements (i.e. mood, theme). However, users also articulate a need to explore music that shares similarities with their favorite tracks or artists, such as \textit{I need a similar track to Superstition by Stevie Wonder}. To address these concerns, this paper proposes an improved Text-to-Music Retrieval model, denoted as TTMR++, which utilizes rich text descriptions generated with a finetuned large language model and metadata. To accomplish this, we obtained various types of seed text from several existing music tag and caption datasets and a knowledge graph dataset of artists and tracks. The experimental results show the effectiveness of TTMR++ in comparison to state-of-the-art music-text joint embedding models through a comprehensive evaluation involving various musical text queries.

comment: Accepted for publication at the IEEE ICASSP 2024

☆ FedMAC: Tackling Partial-Modality Missing in Federated Learning with Cross-Modal Aggregation and Contrastive Regularization

Manh Duong Nguyen, Trung Thanh Nguyen, Huy Hieu Pham, Trong Nghia Hoang, Phi Le Nguyen, Thanh Trung Huynh

Federated Learning (FL) is a method for training machine learning models using distributed data sources. It ensures privacy by allowing clients to collaboratively learn a shared global model while storing their data locally. However, a significant challenge arises when dealing with missing modalities in clients' datasets, where certain features or modalities are unavailable or incomplete, leading to heterogeneous data distribution. While previous studies have addressed the issue of complete-modality missing, they fail to tackle partial-modality missing on account of severe heterogeneity among clients at an instance level, where the pattern of missing data can vary significantly from one sample to another. To tackle this challenge, this study proposes a novel framework named FedMAC, designed to address multi-modality missing under conditions of partial-modality missing in FL. Additionally, to avoid trivial aggregation of multi-modal features, we introduce contrastive-based regularization to impose additional constraints on the latent representation space. The experimental results demonstrate the effectiveness of FedMAC across various client configurations with statistical heterogeneity, outperforming baseline methods by up to 26% in severe missing scenarios, highlighting its potential as a solution for the challenge of partially missing modalities in federated systems.

comment: The 22nd International Symposium on Network Computing and Applications (NCA 2024)

☆ SONIQUE: Video Background Music Generation Using Unpaired Audio-Visual Data

Liqian Zhang, Magdalena Fuentes

We present SONIQUE, a model for generating background music tailored to video content. Unlike traditional video-to-music generation approaches, which rely heavily on paired audio-visual datasets, SONIQUE leverages unpaired data, combining royalty-free music and independent video sources. By utilizing large language models (LLMs) for video understanding and converting visual descriptions into musical tags, alongside a U-Net-based conditional diffusion model, SONIQUE enables customizable music generation. Users can control specific aspects of the music, such as instruments, genres, tempo, and melodies, ensuring the generated output fits their creative vision. SONIQUE is open-source, with a demo available online.

☆ Chain-of-Jailbreak Attack for Image Generation Models via Editing Step by Step

Wenxuan Wang, Kuiyi Gao, Zihan Jia, Youliang Yuan, Jen-tse Huang, Qiuzhi Liu, Shuai Wang, Wenxiang Jiao, Zhaopeng Tu

Text-based image generation models, such as Stable Diffusion and DALL-E 3, hold significant potential in content creation and publishing workflows, making them the focus in recent years. Despite their remarkable capability to generate diverse and vivid images, considerable efforts are being made to prevent the generation of harmful content, such as abusive, violent, or pornographic material. To assess the safety of existing models, we introduce a novel jailbreaking method called Chain-of-Jailbreak (CoJ) attack, which compromises image generation models through a step-by-step editing process. Specifically, for malicious queries that cannot bypass the safeguards with a single prompt, we intentionally decompose the query into multiple sub-queries. The image generation models are then prompted to generate and iteratively edit images based on these sub-queries. To evaluate the effectiveness of our CoJ attack method, we constructed a comprehensive dataset, CoJ-Bench, encompassing nine safety scenarios, three types of editing operations, and three editing elements. Experiments on four widely-used image generation services provided by GPT-4V, GPT-4o, Gemini 1.5 and Gemini 1.5 Pro, demonstrate that our CoJ attack method can successfully bypass the safeguards of models for over 60% cases, which significantly outperforms other jailbreaking methods (i.e., 14%). Further, to enhance these models' safety against our CoJ attack method, we also propose an effective prompting-based method, Think Twice Prompting, that can successfully defend over 95% of CoJ attack. We release our dataset and code to facilitate the AI safety research.

☆ M2AR: A Web-based Modeling Environment for the Augmented Reality Workflow Modeling Language

Fabian Muff, Hans-Georg Fill

This paper introduces M2AR, a new web-based, two- and three-dimensional modeling environment that enables the modeling and execution of augmented reality applications without requiring programming knowledge. The platform is based on a 3D JavaScript library and the mixed reality immersive web standard WebXR. For a first demonstration of its feasibility, the previously introduced Augmented Reality Workflow Modeling Language (ARWFML) has been successfully implemented using this environment. The usefulness of the new modeling environment is demonstrated by showing use cases of the ARWFML on M2AR.

♻ ☆ VideoCLIP-XL: Advancing Long Description Understanding for Video CLIP Models EMNLP 2024

Jiapeng Wang, Chengyu Wang, Kunzhe Huang, Jun Huang, Lianwen Jin

comment: EMNLP 2024 Main conference

♻ ☆ Deepfake Detection: A Comprehensive Survey from the Reliability Perspective

Tianyi Wang, Xin Liao, Kam Pui Chow, Xiaodong Lin, Yinglong Wang

comment: Accepted to ACM Computing Surveys

♻ ☆ TGIF: Text-Guided Inpainting Forgery Dataset

Hannes Mareen, Dimitrios Karageorgiou, Glenn Van Wallendael, Peter Lambert, Symeon Papadopoulos

comment: 6 pages, accepted at IEEE WIFS 2024

Performance 3

☆ AraSync: Precision Time Synchronization in Rural Wireless Living Lab

Md Nadim, Taimoor Ul Islam, Salil Reddy, Tianyi Zhang, Zhibo Meng, Reshal Afzal, Sarath Babu, Arsalan Ahmad, Daji Qiao, Anish Arora, Hongwei Zhang

Time synchronization is a critical component in network operation and management, and it is also required by Ultra-Reliable, Low-Latency Communications (URLLC) in next-generation wireless systems such as those of 5G, 6G, and Open RAN. In this context, we design and implement AraSync as an end-to-end time synchronization system in the ARA wireless living lab to enable advanced wireless experiments and applications involving stringent time constraints. We make use of Precision Time Protocol (PTP) at different levels to achieve synchronization accuracy in the order of nanoseconds. Along with fiber networks, AraSync enables time synchronization across the AraHaul wireless x-haul network consisting of long-range, high-capacity mmWave and microwave links. In this paper, we present the detailed design and implementation of AraSync, including its hardware and software components and the PTP network topology. Further, we experimentally characterize the performance of AraSync from spatial and temporal dimensions. Our measurement and analysis of the clock offset and mean path delay show the impact of the wireless channel and weather conditions on the PTP synchronization accuracy.

comment: 8 pages, 10 figures, accepted in ACM WiNTECH 2024 (The 18th ACM Workshop on Wireless Network Testbeds, Experimental evaluation & Characterization 2024)

☆ EXAQ: Exponent Aware Quantization For LLMs Acceleration

Moran Shkolnik, Maxim Fishman, Brian Chmiel, Hilla Ben-Yaacov, Ron Banner, Kfir Yehuda Levy

Quantization has established itself as the primary approach for decreasing the computational and storage expenses associated with Large Language Models (LLMs) inference. The majority of current research emphasizes quantizing weights and activations to enable low-bit general-matrix-multiply (GEMM) operations, with the remaining non-linear operations executed at higher precision. In our study, we discovered that following the application of these techniques, the primary bottleneck in LLMs inference lies in the softmax layer. The softmax operation comprises three phases: exponent calculation, accumulation, and normalization, Our work focuses on optimizing the first two phases. We propose an analytical approach to determine the optimal clipping value for the input to the softmax function, enabling sub-4-bit quantization for LLMs inference. This method accelerates the calculations of both $e^x$ and $\sum(e^x)$ with minimal to no accuracy degradation. For example, in LLaMA1-30B, we achieve baseline performance with 2-bit quantization on the well-known "Physical Interaction: Question Answering" (PIQA) dataset evaluation. This ultra-low bit quantization allows, for the first time, an acceleration of approximately 4x in the accumulation phase. The combination of accelerating both $e^x$ and $\sum(e^x)$ results in a 36.9% acceleration in the softmax operation.

♻ ☆ Scaler: Efficient and Effective Cross Flow Analysis

Steven, Tang, Mingcan Xiang, Yang Wang, Bo Wu, Jianjun Chen, Tongping Liu

Performance analysis is challenging as different components (e.g.,different libraries, and applications) of a complex system can interact with each other. However, few existing tools focus on understanding such interactions. To bridge this gap, we propose a novel analysis method "Cross Flow Analysis (XFA)" that monitors the interactions/flows across these components. We also built the Scaler profiler that provides a holistic view of the time spent on each component (e.g., library or application) and every API inside each component. This paper proposes multiple new techniques, such as Universal Shadow Table, and Relation-Aware Data Folding. These techniques enable Scaler to achieve low runtime overhead, low memory overhead, and high profiling accuracy. Based on our extensive experimental results, Scaler detects multiple unknown performance issues inside widely-used applications, and therefore will be a useful complement to existing work. The reproduction package including the source code, benchmarks, and evaluation scripts, can be found at https://doi.org/10.5281/zenodo.13336658.

comment: Paper has been accepted by ASE'24 https://conf.researchr.org/details/ase-2024/ase-2024-research/73/Scaler-Efficient-and-Effective-Cross-Flow-Analysis

Distributed, Parallel, and Cluster Computing 9

☆ FedStein: Enhancing Multi-Domain Federated Learning Through James-Stein Estimator NeurIPS 2024

Sunny Gupta, Nikita Jangid, Amit Sethi

comment: 12 pages, 2 figures. Accepted at International Workshop on Federated Foundation Models In Conjunction with NeurIPS 2024 (FL@FM-NeurIPS'24)

☆ SeBS-Flow: Benchmarking Serverless Cloud Function Workflows

Larissa Schmid, Marcin Copik, Alexandru Calotoiu, Laurin Brandner, Anne Koziolek, Torsten Hoefler

☆ Influence-oriented Personalized Federated Learning

Yue Tan, Guodong Long, Jing Jiang, Chengqi Zhang

☆ Memory-distributed level set-based inverse homogenisation of three-dimensional piezoelectric materials

Zachary J. Wegert, Anthony P. Roberts, Vivien J. Challis

In this paper we use level set-based topology optimisation to design three-dimensional periodic piezoelectric materials with enhanced properties. Our methodology is fully memory-distributed and written in Julia using the package GridapTopOpt. We compare and assess several existing iterative solvers with respect to their weak scalability and find that an approximate Schur complement preconditioned GMRES method demonstrates the best performance and scalability for solving the piezoelectric homogenisation equations. We use the developed techniques to computationally design high-resolution piezoelectric metamaterials with enhanced stiffness and piezoelectric properties that yield new insights into material design for sensor, hydrophone, and actuator applications. We suggest two robust structures with simple geometric features that exhibit enhanced piezoelectric properties several times larger than those of the base material. We find that level set-based topology optimisation is well suited to problems involving piezoelectricity and has the advantage of avoiding large regions of intermediate density material.

☆ FedCert: Federated Accuracy Certification

Minh Hieu Nguyen, Huu Tien Nguyen, Trung Thanh Nguyen, Manh Duong Nguyen, Trong Nghia Hoang, Truong Thao Nguyen, Phi Le Nguyen

Federated Learning (FL) has emerged as a powerful paradigm for training machine learning models in a decentralized manner, preserving data privacy by keeping local data on clients. However, evaluating the robustness of these models against data perturbations on clients remains a significant challenge. Previous studies have assessed the effectiveness of models in centralized training based on certified accuracy, which guarantees that a certain percentage of the model's predictions will remain correct even if the input data is perturbed. However, the challenge of extending these evaluations to FL remains unresolved due to the unknown client's local data. To tackle this challenge, this study proposed a method named FedCert to take the first step toward evaluating the robustness of FL systems. The proposed method is designed to approximate the certified accuracy of a global model based on the certified accuracy and class distribution of each client. Additionally, considering the Non-Independent and Identically Distributed (Non-IID) nature of data in real-world scenarios, we introduce the client grouping algorithm to ensure reliable certified accuracy during the aggregation step of the approximation algorithm. Through theoretical analysis, we demonstrate the effectiveness of FedCert in assessing the robustness and reliability of FL systems. Moreover, experimental results on the CIFAR-10 and CIFAR-100 datasets under various scenarios show that FedCert consistently reduces the estimation error compared to baseline methods. This study offers a solution for evaluating the robustness of FL systems and lays the groundwork for future research to enhance the dependability of decentralized learning. The source code is available at https://github.com/thanhhff/FedCert/.

comment: The 22nd International Symposium on Network Computing and Applications (NCA 2024)

♻ ☆ TOB-SVD: Total-Order Broadcast with Single-Vote Decisions in the Sleepy Model

Francesco D'Amato, Roberto Saltini, Thanh-Hai Tran, Luca Zanolini

Over the past years, distributed consensus research has extended its focus towards addressing challenges in large-scale, permissionless systems, such as blockchains. This shift is characterized by the need to accommodate dynamic participation, contrasting the traditional approach of a static set of continuously online participants. Works like Bitcoin and the sleepy model have set the stage for this evolving framework. Notable contributions from Momose and Ren (CCS 2022) and subsequent works have introduced Total-Order Broadcast protocols leveraging Graded Agreement primitives and supporting dynamic participation. However, these approaches often require multiple phases of voting per decision, creating a potential bottleneck for real-world large-scale systems. Addressing this, our paper introduces TOB-SVD, a novel Total-Order Broadcast protocol in the sleepy model, which is resilient to up to 1/2 of adversarial participants. TOB-SVD requires only a single phase of voting per decision in the best case and achieves lower expected latency compared to existing approaches offering the same optimal adversarial resilience. This work paves the way to more practical Total-Order Broadcast protocols to be implemented in real-world systems where a large number of participants are involved simultaneously and their participation level might fluctuate over time.

♻ ☆ Local certification of forbidden subgraphs

Nicolas Bousquet, Linda Cook, Laurent Feuilloley, Théo Pierron, Sébastien Zeitoun

Detecting specific structures in a network has been a very active theme of research in distributed computing for at least a decade. In this paper, we start the study of subgraph detection from the perspective of local certification. Remember that a local certification is a distributed mechanism enabling the nodes of a network to check the correctness of the current configuration, thanks to small pieces of information called certificates. Our main question is: For a given graph $H$, what is the minimum certificate size that allows checking that the network does not contain $H$ as a (possibly induced) subgraph? We show a variety of lower and upper bounds, uncovering an interesting interplay between the optimal certificate size, the size of the forbidden subgraph, and the locality of the verification. Along the way we introduce several new technical tools, in particular what we call the \emph{layered map}, which is not specific to forbidden subgraphs and that we expect to be useful for certifying many other properties.

♻ ☆ Scattered Mixture-of-Experts Implementation

Shawn Tan, Yikang Shen, Rameswar Panda, Aaron Courville

We present ScatterMoE, an implementation of Sparse Mixture-of-Experts (SMoE) on GPUs. ScatterMoE builds upon existing implementations, and overcoming some of the limitations to improve inference and training speed, and memory footprint. This implementation achieves this by avoiding padding and making excessive copies of the input. We introduce ParallelLinear, the main component we use to build our implementation and the various kernels used to speed up the operation. We benchmark our implementation against Megablocks, and show that it enables a higher throughput and lower memory footprint. We also show how ParallelLinear enables extension of the Mixture-of-Experts concept by demonstrating with an implementation of Mixture of Attention.

♻ ☆ Generic Multicast (Extended Version)

José Augusto Bolina, Pierre Sutra, Douglas Antunes Rocha, Lasaro Camargos

Communication primitives play a central role in modern computing. They offer a panel of reliability and ordering guarantees for messages, enabling the implementation of complex distributed interactions. In particular, atomic broadcast is a pivotal abstraction for implementing fault-tolerant distributed services. This primitive allows disseminating messages across the system in a total order. There are two group communication primitives closely related to atomic broadcast. Atomic multicast permits targeting a subset of participants, possibly stricter than the whole system. Generic broadcast leverages the semantics of messages to order them only where necessary (that is when they conflict). In this paper, we propose to combine all these primitives into a single, more general one, called generic multicast. We formally specify the guarantees offered by generic multicast and present efficient algorithms. Compared to prior works, our solutions offer appealing properties in terms of time and space complexity. In particular, when a run is conflict-free, that is no two messages conflict, a message is delivered after at most three message delays.

comment: 12 pages, 2 figures

2024-10-03

Hardware Architecture 4

☆ Custom Non-Linear Model Predictive Control for Obstacle Avoidance in Indoor and Outdoor Environments

Lara Laban, Mariusz Wzorek, Piotr Rudol, Tommy Persson

Navigating complex environments requires Unmanned Aerial Vehicles (UAVs) and autonomous systems to perform trajectory tracking and obstacle avoidance in real-time. While many control strategies have effectively utilized linear approximations, addressing the non-linear dynamics of UAV, especially in obstacle-dense environments, remains a key challenge that requires further research. This paper introduces a Non-linear Model Predictive Control (NMPC) framework for the DJI Matrice 100, addressing these challenges by using a dynamic model and B-spline interpolation for smooth reference trajectories, ensuring minimal deviation while respecting safety constraints. The framework supports various trajectory types and employs a penalty-based cost function for control accuracy in tight maneuvers. The framework utilizes CasADi for efficient real-time optimization, enabling the UAV to maintain robust operation even under tight computational constraints. Simulation and real-world indoor and outdoor experiments demonstrated the NMPC ability to adapt to disturbances, resulting in smooth, collision-free navigation.

comment: This manuscript has 7 pages and 8 figures, detailing NMPC for UAV obstacle avoidance using DJI UAVs. It features simulations, experimental results, and uses CasADi for optimization with ROS integration. Code and media at https://github.com/larasupernovae/nmpc_flash_multi_obstacle

☆ AiBAT: Artificial Intelligence/Instructions for Build, Assembly, and Test

Benjamin Nuernberger, Anny Liu, Heather Stefanini, Richard Otis, Amanda Towler, R. Peter Dillon

Instructions for Build, Assembly, and Test (IBAT) refers to the process used whenever any operation is conducted on hardware, including tests, assembly, and maintenance. Currently, the generation of IBAT documents is time-intensive, as users must manually reference and transfer information from engineering diagrams and parts lists into IBAT instructions. With advances in machine learning and computer vision, however, it is possible to have an artificial intelligence (AI) model perform the partial filling of the IBAT template, freeing up engineer time for more highly skilled tasks. AiBAT is a novel system for assisting users in authoring IBATs. It works by first analyzing assembly drawing documents, extracting information and parsing it, and then filling in IBAT templates with the extracted information. Such assisted authoring has potential to save time and reduce cost. This paper presents an overview of the AiBAT system, including promising preliminary results and discussion on future work.

comment: 9 pages, 6 figures, 2 tables

♻ ☆ Modeling the Energy Consumption of the HEVC Software Encoding Process using Processor events

Geetha Ramasubbu, Andrè Kaup, Christian Herglotz

Developing energy-efficient video encoding algorithms is highly important due to the high processing complexities and, consequently, the high energy demand of the encoding process. To accomplish this, the energy consumption of the video encoders must be studied, which is only possible with a complex and dedicated energy measurement setup. This emphasizes the need for simple energy estimation models, which estimate the energy required for the encoding. Our paper investigates the possibility of estimating the energy demand of a HEVC software CPU-encoding process using processor events. First, we perform energy measurements and obtain processor events using dedicated profiling software. Then, by using the measured energy demand of the encoding process and profiling data, we build an encoding energy estimation model that uses the processor events of the ultrafast encoding preset to obtain the energy estimate for complex encoding presets with a mean absolute percentage error of 5.36% when averaged over all the presets. Additionally, we present an energy model that offers the possibility of obtaining energy distribution among various encoding sub-processes.

♻ ☆ GPUDrive: Data-driven, multi-agent driving simulation at 1 million FPS

Saman Kazemkhani, Aarav Pandya, Daphne Cornelisse, Brennan Shacklett, Eugene Vinitsky

Multi-agent learning algorithms have been successful at generating superhuman planning in various games but have had limited impact on the design of deployed multi-agent planners. A key bottleneck in applying these techniques to multi-agent planning is that they require billions of steps of experience. To enable the study of multi-agent planning at scale, we present GPUDrive, a GPU-accelerated, multi-agent simulator built on top of the Madrona Game Engine that can generate over a million simulation steps per second. Observation, reward, and dynamics functions are written directly in C++, allowing users to define complex, heterogeneous agent behaviors that are lowered to high-performance CUDA. We show that using GPUDrive we can effectively train reinforcement learning agents over many scenes in the Waymo Open Motion Dataset, yielding highly effective goal-reaching agents in minutes for individual scenes and enabling agents to navigate thousands of scenarios within hours. The code base with pre-trained agents is available at \url{https://github.com/Emerge-Lab/gpudrive}.

Distributed, Parallel, and Cluster Computing 18

☆ EinDecomp: Decomposition of Declaratively-Specified Machine Learning and Numerical Computations for Parallel Execution

Daniel Bourgeois, Zhimin Ding, Dimitrije Jankov, Jiehui Li, Mahmoud Sleem, Yuxin Tang, Jiawen Yao, Xinyu Yao, Chris Jermaine

We consider the problem of automatically decomposing operations over tensors or arrays so that they can be executed in parallel on multiple devices. We address two, closely-linked questions. First, what programming abstraction should systems for tensor-based computing offer to enable such decompositions? Second, given that abstraction, how should such systems automatically decompose a tensor-based computation? We assert that tensor-based systems should offer a programming abstraction based on an extended Einstein summation notation, which is a fully declarative, mathematical specification for tensor computations. We show that any computation specified in the Einstein summation notation can be re-written into an equivalent tensor-relational computation, and this re-write generalizes existing notations of tensor parallelism such as "data parallel'' and "model parallel.'' We consider the algorithmic problem of optimally computing a tensor-relational decomposition of a graph of operations specified in our extended Einstein summation notation, and we experimentally show the value of the algorithm that we develop.

☆ Disaggregated Memory with SmartNIC Offloading: a Case Study on Graph Processing

Jacob Wahlgren, Gabin Schieffer, Maya Gokhale, Roger Pearce, Ivy Peng

Disaggregated memory breaks the boundary of monolithic servers to enable memory provisioning on demand. Using network-attached memory to provide memory expansion for memory-intensive applications on compute nodes can improve the overall memory utilization on a cluster and reduce the total cost of ownership. However, current software solutions for leveraging network-attached memory must consume resources on the compute node for memory management tasks. Emerging off-path smartNICs provide general-purpose programmability at low-cost low-power cores. This work provides a general architecture design that enables network-attached memory and offloading tasks onto off-path programmable SmartNIC. We provide a prototype implementation called SODA on Nvidia BlueField DPU. SODA adapts communication paths and data transfer alternatives, pipelines data movement stages, and enables customizable data caching and prefetching optimizations. We evaluate SODA in five representative graph applications on real-world graphs. Our results show that SODA can achieve up to 7.9x speedup compared to node-local SSD and reduce network traffic by 42% compared to disaggregated memory without SmartNIC offloading at similar or better performance.

☆ Federated k-Core Decomposition: A Secure Distributed Approach

Bin Guo, Emil Sekerinski, Lingyang Chu

As one of the most well-studied cohesive subgraph models, the $k$-core is widely used to find graph nodes that are ``central'' or ``important'' in many applications, such as biological networks, social networks, ecological networks, and financial networks. For distributed networks, e.g., Decentralized Online Social Networks (DOSNs) such that each vertex is a client as a single computing unit, the distributed $k$-core decomposition algorithms are already proposed. However, current distributed approaches fail to adequately protect privacy and security. In today's data-driven world, data privacy and security have attracted more and more attention, e.g., DOSNs are proposed to protect privacy by storing user information locally without using a single centralized server. In this work, we are the first to propose the secure version of the distributed $k$-core decomposition.

comment: 14 pages, 7 figures

☆ Fair Decentralized Learning

Sayan Biswas, Anne-Marie Kermarrec, Rishi Sharma, Thibaud Trinca, Martijn de Vos

Decentralized learning (DL) is an emerging approach that enables nodes to collaboratively train a machine learning model without sharing raw data. In many application domains, such as healthcare, this approach faces challenges due to the high level of heterogeneity in the training data's feature space. Such feature heterogeneity lowers model utility and negatively impacts fairness, particularly for nodes with under-represented training data. In this paper, we introduce \textsc{Facade}, a clustering-based DL algorithm specifically designed for fair model training when the training data exhibits several distinct features. The challenge of \textsc{Facade} is to assign nodes to clusters, one for each feature, based on the similarity in the features of their local data, without requiring individual nodes to know apriori which cluster they belong to. \textsc{Facade} (1) dynamically assigns nodes to their appropriate clusters over time, and (2) enables nodes to collaboratively train a specialized model for each cluster in a fully decentralized manner. We theoretically prove the convergence of \textsc{Facade}, implement our algorithm, and compare it against three state-of-the-art baselines. Our experimental results on three datasets demonstrate the superiority of our approach in terms of model accuracy and fairness compared to all three competitors. Compared to the best-performing baseline, \textsc{Facade} on the CIFAR-10 dataset also reduces communication costs by 32.3\% to reach a target accuracy when cluster sizes are imbalanced.

☆ An Edge-Computing based Industrial Gateway for Industry 4.0 using ARM TrustZone Technology

Sandeep Gupta

Secure and efficient communication to establish a seamless nexus between the five levels of a typical automation pyramid is paramount to Industry 4.0. Specifically, vertical and horizontal integration of these levels is an overarching requirement to accelerate productivity and improve operational activities. Vertical integration can improve visibility, flexibility, and productivity by connecting systems and applications. Horizontal integration can provide better collaboration and adaptability by connecting internal production facilities, multi-site operations, and third-party partners in a supply chain. In this paper, we propose an Edge-computing-based Industrial Gateway for interfacing information technology and operational technology that can enable Industry 4.0 vertical and horizontal integration. Subsequently, we design and develop a working prototype to demonstrate a remote production-line maintenance use case with a strong focus on security aspects and the edge paradigm to bring computational resources and data storage closer to data sources.

☆ Temporal Predictive Coding for Gradient Compression in Distributed Learning

Adrian Edin, Zheng Chen, Michel Kieffer, Mikael Johansson

This paper proposes a prediction-based gradient compression method for distributed learning with event-triggered communication. Our goal is to reduce the amount of information transmitted from the distributed agents to the parameter server by exploiting temporal correlation in the local gradients. We use a linear predictor that \textit{combines past gradients to form a prediction of the current gradient}, with coefficients that are optimized by solving a least-square problem. In each iteration, every agent transmits the predictor coefficients to the server such that the predicted local gradient can be computed. The difference between the true local gradient and the predicted one, termed the \textit{prediction residual, is only transmitted when its norm is above some threshold.} When this additional communication step is omitted, the server uses the prediction as the estimated gradient. This proposed design shows notable performance gains compared to existing methods in the literature, achieving convergence with reduced communication costs.

comment: 8 pages, 3 figures, presented at the 60th Allerton conference on Communication, Control, and Computing

☆ Personalized Federated Learning for Generative AI-Assisted Semantic Communications

Yubo Peng, Feibo Jiang, Li Dong, Kezhi Wang, Kun Yang

Semantic Communication (SC) focuses on transmitting only the semantic information rather than the raw data. This approach offers an efficient solution to the issue of spectrum resource utilization caused by the various intelligent applications on Mobile Users (MUs). Generative Artificial Intelligence (GAI) models have recently exhibited remarkable content generation and signal processing capabilities, presenting new opportunities for enhancing SC. Therefore, we propose a GAI-assisted SC (GSC) model deployed between MUs and the Base Station (BS). Then, to train the GSC model using the local data of MUs while ensuring privacy and accommodating heterogeneous requirements of MUs, we introduce Personalized Semantic Federated Learning (PSFL). This approach incorporates a novel Personalized Local Distillation (PLD) and Adaptive Global Pruning (AGP). In PLD, each MU selects a personalized GSC model as a mentor tailored to its local resources and a unified Convolutional Neural Networks (CNN)-based SC (CSC) model as a student. This mentor model is then distilled into the student model for global aggregation. In AGP, we perform network pruning on the aggregated global model according to real-time communication environments, reducing communication energy. Finally, numerical results demonstrate the feasibility and efficiency of the proposed PSFL scheme.

☆ LLM-Pilot: Characterize and Optimize Performance of your LLM Inference Services SC '24

Małgorzata Łazuka, Andreea Anghel, Thomas Parnell

As Large Language Models (LLMs) are rapidly growing in popularity, LLM inference services must be able to serve requests from thousands of users while satisfying performance requirements. The performance of an LLM inference service is largely determined by the hardware onto which it is deployed, but understanding of which hardware will deliver on performance requirements remains challenging. In this work we present LLM-Pilot - a first-of-its-kind system for characterizing and predicting performance of LLM inference services. LLM-Pilot performs benchmarking of LLM inference services, under a realistic workload, across a variety of GPUs, and optimizes the service configuration for each considered GPU to maximize performance. Finally, using this characterization data, LLM-Pilot learns a predictive model, which can be used to recommend the most cost-effective hardware for a previously unseen LLM. Compared to existing methods, LLM-Pilot can deliver on performance requirements 33% more frequently, whilst reducing costs by 60% on average.

comment: Accepted to the International Conference for High Performance Computing, Networking, Storage and Analysis (SC '24)

☆ Model-guided Fuzzing of Distributed Systems

Ege Berkay Gulcan, Burcu Kulahcioglu Ozkan, Rupak Majumdar, Srinidhi Nagendra

☆ Selection Guidelines for Geographical SMR Protocols: A Communication Pattern-based Latency Modeling Approach

Kohya Shiozaki, Junya Nakamura

State machine replication (SMR) is a replication technique that ensures fault tolerance by duplicating a service. Geographical SMR can enhance its robustness against disasters by distributing replicas in separate geographical locations. Several geographical SMR protocols have been proposed in the literature, each of which tailored to specific requirements; for example, protocols designed to meet the requirement of latency reduction by either sacrificing a part of their fault tolerance or limiting the content of responses to clients. However, this diversity complicates the decision-making process for selecting the best protocol for a particular service. In this study, we introduce a latency estimation model for these SMR protocols based on the communication patterns of the protocols and perform simulations for various cases. Based on the simulation results and an experimental evaluation, we present five selection guidelines for geographical SMR protocols based on their log management policy, distances between replicas, number of replicas, frequency of slow paths, and client distribution. These selection guidelines enable determining the best geographical SMR protocol for each situation.

comment: This paper was submitted to the 26th International Symposium on Stabilization, Safety, and Security of Distributed Systems

☆ MTDNS: Moving Target Defense for Resilient DNS Infrastructure

Abdullah Aydeger, Pei Zhou, Sanzida Hoque, Marco Carvalho, Engin Zeydan

One of the most critical components of the Internet that an attacker could exploit is the DNS (Domain Name System) protocol and infrastructure. Researchers have been constantly developing methods to detect and defend against the attacks against DNS, specifically DNS flooding attacks. However, most solutions discard packets for defensive approaches, which can cause legitimate packets to be dropped, making them highly dependable on detection strategies. In this paper, we propose MTDNS, a resilient MTD-based approach that employs Moving Target Defense techniques through Software Defined Networking (SDN) switches to redirect traffic to alternate DNS servers that are dynamically created and run under the Network Function Virtualization (NFV) framework. The proposed approach is implemented in a testbed environment by running our DNS servers as separate Virtual Network Functions, NFV Manager, SDN switches, and an SDN Controller. The experimental result shows that the MTDNS approach achieves a much higher success rate in resolving DNS queries and significantly reduces average latency even if there is a DNS flooding attack.

comment: 6 pages, Accepted for publication at IEEE CCNC 2025

☆ Extracting the Potential of Emerging Hardware Accelerators for Symmetric Eigenvalue Decomposition

Hansheng Wang, Lu Shi, Zhekai duan, Panruo Wu, Liwei Guo, Shaoshuai Zhang

Benefiting from the advancement of hardware accelerators such as GPUs, deep neural networks and scientific computing applications can achieve superior performance. Recently, the computing capacity of emerging hardware accelerators has increased rapidly, while memory bandwidth has not kept pace with this growth. This disparity exacerbates the gap between computing and memory, leading to inefficiencies on conventional algorithms, as they're likely to be converted from compute-bound to memory-bound. Symmetric eigenvalue decomposition (EVD), a critical operation in various research domains including scientific computing, deep learning training, and inference algorithms, exhibits suboptimal performance due to achieving less than 3\% hardware computing utilization on the H100 GPU. In this paper, we analyze the features of emerging hardware accelerators to identify the bottlenecks inherent in conventional EVD algorithms. To improve EVD performance, we propose several algorithmic optimizations aimed at solving the memory-bound problem and providing a better utilization of the rich computing capacity and parallelism on the emerging hardware accelerators. Experimentally, our proposed method demonstrates significant speedups on tridiagonalization, which is the main workload that takes over 90\% elapsed time of EVD, compared to the SOTA cuSOLVER tridiagonalization, achieving up to 10.1x, 7.5x, and 2.3x improvements on H100, A100, and RTX 4090 GPUs, respectively. And the end-to-end the performance of EVD solver is also up to 4.1x faster than cuSOVLER.

☆ FedPeWS: Personalized Warmup via Subnetworks for Enhanced Heterogeneous Federated Learning

Nurbek Tastan, Samuel Horvath, Martin Takac, Karthik Nandakumar

Statistical data heterogeneity is a significant barrier to convergence in federated learning (FL). While prior work has advanced heterogeneous FL through better optimization objectives, these methods fall short when there is extreme data heterogeneity among collaborating participants. We hypothesize that convergence under extreme data heterogeneity is primarily hindered due to the aggregation of conflicting updates from the participants in the initial collaboration rounds. To overcome this problem, we propose a warmup phase where each participant learns a personalized mask and updates only a subnetwork of the full model. This personalized warmup allows the participants to focus initially on learning specific subnetworks tailored to the heterogeneity of their data. After the warmup phase, the participants revert to standard federated optimization, where all parameters are communicated. We empirically demonstrate that the proposed personalized warmup via subnetworks (FedPeWS) approach improves accuracy and convergence speed over standard federated optimization methods.

♻ ☆ Preble: Efficient Distributed Prompt Scheduling for LLM Serving

Vikranth Srivatsa, Zijian He, Reyna Abhyankar, Dongming Li, Yiying Zhang

Prompts to large language models (LLMs) have evolved beyond simple user questions. For LLMs to solve complex problems, today's practices are to include domain-specific instructions, illustration of tool usages, and/or long context such as textbook chapters in prompts. As such, many parts of prompts are repetitive across requests. Recent works propose to cache and reuse KV state of prompts. However, they are all confined to a single-GPU optimization, while production LLM serving systems are distributed by nature. This paper proposes Preble, the first distributed LLM serving platform that targets and optimizes for prompt sharing. We designed a distributed scheduling system that co-optimizes KV state reuse and computation load-balancing with a new scheduling algorithm and a hierarchical scheduling mechanism. Our evaluation of Preble with real workloads and request arrival patterns on two open-source LLMs shows that Preble outperforms the SOTA serving systems by 1.5X to 14.5X on average latency and 2X to 10X on p99 latency.

♻ ☆ TURNIP: A "Nondeterministic" GPU Runtime with CPU RAM Offload

Zhimin Ding, Jiawen Yao, Brianna Barrow, Tania Lorido Botran, Christopher Jermaine, Yuxin Tang, Jiehui Li, Xinyu Yao, Sleem Mahmoud Abdelghafar, Daniel Bourgeois

An obvious way to alleviate memory difficulties in GPU-based AI computing is via CPU offload, where data are moved between GPU and CPU RAM, so inexpensive CPU RAM is used to increase the amount of storage available. While CPU offload is an obvious idea, it can greatly slow down a computation, due to the relatively slow transfer rate between CPU RAM and GPU RAM. Thus, any system for CPU offload needs to ensure that when such a transfer needs to happen, no computation is blocked waiting for the transfer to finish. One of the key challenges when using CPU offload is that memory transfers introduce nondeterminacy into the system: it is not possible to know before runtime when the transfers will finish, and hence what is the best order of operations to run to ensure there is no blocking. In this paper, we describe TURNIP, which is a system for running AI computations using CPU offload. The key innovation in TURNIP is the compilation of the AI computation into a dependency graph that gives the TURNIP runtime freedom to run operations such as GPU kernel calls in many different orders; at runtime, TURNIP chooses the best order in response to real-time events.

♻ ☆ Fast Matrix Multiplications for Lookup Table-Quantized LLMs EMNLP 2024

Han Guo, William Brandon, Radostin Cholakov, Jonathan Ragan-Kelley, Eric P. Xing, Yoon Kim

The deployment of large language models (LLMs) is often constrained by memory bandwidth, where the primary bottleneck is the cost of transferring model parameters from the GPU's global memory to its registers. When coupled with custom kernels that fuse the dequantization and matmul operations, weight-only quantization can thus enable faster inference by reducing the amount of memory movement. However, developing high-performance kernels for weight-quantized LLMs presents substantial challenges, especially when the weights are compressed to non-evenly-divisible bit widths (e.g., 3 bits) with non-uniform, lookup table (LUT) quantization. This paper describes FLUTE, a flexible lookup table engine for LUT-quantized LLMs, which uses offline restructuring of the quantized weight matrix to minimize bit manipulations associated with unpacking, and vectorization and duplication of the lookup table to mitigate shared memory bandwidth constraints. At batch sizes < 32 and quantization group size of 128 (typical in LLM inference), the FLUTE kernel can be 2-4x faster than existing GEMM kernels. As an application of FLUTE, we explore a simple extension to lookup table-based NormalFloat quantization and apply it to quantize LLaMA3 to various configurations, obtaining competitive quantization performance against strong baselines while obtaining an end-to-end throughput increase of 1.5 to 2 times.

comment: EMNLP 2024 (Findings)

♻ ☆ Effective Heterogeneous Federated Learning via Efficient Hypernetwork-based Weight Generation

Yujin Shin, Kichang Lee, Sungmin Lee, You Rim Choi, Hyung-Sin Kim, JeongGil Ko

While federated learning leverages distributed client resources, it faces challenges due to heterogeneous client capabilities. This necessitates allocating models suited to clients' resources and careful parameter aggregation to accommodate this heterogeneity. We propose HypeMeFed, a novel federated learning framework for supporting client heterogeneity by combining a multi-exit network architecture with hypernetwork-based model weight generation. This approach aligns the feature spaces of heterogeneous model layers and resolves per-layer information disparity during weight aggregation. To practically realize HypeMeFed, we also propose a low-rank factorization approach to minimize computation and memory overhead associated with hypernetworks. Our evaluations on a real-world heterogeneous device testbed indicate that \system enhances accuracy by 5.12% over FedAvg, reduces the hypernetwork memory requirements by 98.22%, and accelerates its operations by 1.86x compared to a naive hypernetwork approach. These results demonstrate HypeMeFed's effectiveness in leveraging and engaging heterogeneous clients for federated learning.

♻ ☆ Stalactite: Toolbox for Fast Prototyping of Vertical Federated Learning Systems

Anastasiia Zakharova, Dmitriy Alexandrov, Maria Khodorchenko, Nikolay Butakov, Alexey Vasilev, Maxim Savchenko, Alexander Grigorievskiy

Machine learning (ML) models trained on datasets owned by different organizations and physically located in remote databases offer benefits in many real-world use cases. State regulations or business requirements often prevent data transfer to a central location, making it difficult to utilize standard machine learning algorithms. Federated Learning (FL) is a technique that enables models to learn from distributed datasets without revealing the original data. Vertical Federated learning (VFL) is a type of FL where data samples are divided by features across several data owners. For instance, in a recommendation task, a user can interact with various sets of items, and the logs of these interactions are stored by different organizations. In this demo paper, we present \emph{Stalactite} - an open-source framework for VFL that provides the necessary functionality for building prototypes of VFL systems. It has several advantages over the existing frameworks. In particular, it allows researchers to focus on the algorithmic side rather than engineering and to easily deploy learning in a distributed environment. It implements several VFL algorithms and has a built-in homomorphic encryption layer. We demonstrate its use on a real-world recommendation datasets.

Operating Systems 2

♻ ☆ The eBPF Runtime in the Linux Kernel

Bolaji Gbadamosi, Luigi Leonardi, Tobias Pulls, Toke Høiland-Jørgensen, Simone Ferlin-Reiter, Simo Sorce, Anna Brunström

Extended Berkeley Packet Filter (eBPF) is a runtime that enables users to load programs into the operating system (OS) kernel, like Linux or Windows, and execute them safely and efficiently at designated kernel hooks. Each program passes through a verifier that reasons about the safety guarantees for execution. Hosting a safe virtual machine runtime within the kernel makes it dynamically programmable. Unlike the popular approach of bypassing or completely replacing the kernel, eBPF gives users the flexibility to modify the kernel on the fly, rapidly experiment and iterate, and deploy solutions to achieve their workload-specific needs, while working in concert with the kernel. In this paper, we present the first comprehensive description of the design and implementation of the eBPF runtime in the Linux kernel. We argue that eBPF today provides a mature and safe programming environment for the kernel. It has seen wide adoption since its inception and is increasingly being used not just to extend, but program entire components of the kernel, while preserving its runtime integrity. We outline the compelling advantages it offers for real-world production usage, and illustrate current use cases. Finally, we identify its key challenges, and discuss possible future directions.

comment: 22 pages, 6 figures

♻ ☆ Securing Cloud File Systems with Trusted Execution

Quinn Burke, Yohan Beugin, Blaine Hoak, Rachel King, Eric Pauley, Ryan Sheatsley, Mingli Yu, Ting He, Thomas La Porta, Patrick McDaniel

Cloud file systems offer organizations a scalable and reliable file storage solution. However, cloud file systems have become prime targets for adversaries, and traditional designs are not equipped to protect organizations against the myriad of attacks that may be initiated by a malicious cloud provider, co-tenant, or end-client. Recently proposed designs leveraging cryptographic techniques and trusted execution environments (TEEs) still force organizations to make undesirable trade-offs, consequently leading to either security, functional, or performance limitations. In this paper, we introduce BFS, a cloud file system that leverages the security capabilities provided by TEEs to bootstrap new security protocols that deliver strong security guarantees, high-performance, and a transparent POSIX-like interface to clients. BFS delivers stronger security guarantees and up to a 2.5X speedup over a state-of-the-art secure file system. Moreover, compared to the industry standard NFS, BFS achieves up to 2.2X speedups across micro-benchmarks and incurs <1X overhead for most macro-benchmark workloads. BFS demonstrates a holistic cloud file system design that does not sacrifice an organizations' security yet can embrace all of the functional and performance advantages of outsourcing.

Programming Languages 3

☆ The Long Way to Deforestation (Technical Report): A Type Inference and Elaboration Technique for Removing Intermediate Data Structures

Yijia Chen, Lionel Parreaux

Deforestation is a compiler optimization that removes intermediate data structure allocations from functional programs to improve their efficiency. This is an old idea, but previous approaches have proved limited or impractical: they either only worked on compositions of predefined combinators (shortcut fusion), or involved the aggressive unfolding of recursive definitions until a depth limit was reached or a reoccurring pattern was found to tie the recursive knot, resulting in impractical algorithmic complexity and large amounts of code duplication. We present Lumberhack, a general-purpose deforestation approach for purely functional call-by-value programs. Lumberhack uses subtype inference to reason about data structure production and consumption and uses an elaboration pass to fuse the corresponding recursive definitions. It fuses large classes of mutually recursive definitions while avoiding much of the unproductive (and sometimes counter-productive) code duplication inherent in previous approaches. We prove the soundness of Lumberhack using logical relations and experimentally demonstrate significant speedups in the standard nofib benchmark suite.

comment: This is the technical report version of the paper published at ICFP 2024

☆ The why, what, and how of AI-based coding in scientific research

Tonghe Zhuang, Zhicheng Lin

Computer programming (coding) is indispensable for researchers across disciplines, yet it remains challenging to learn and time-consuming to carry out. Generative AI, particularly large language models (LLMs), has the potential to transform coding into intuitive conversations, but best practices and effective workflows are only emerging. We dissect AI-based coding through three key lenses: the nature and role of LLMs in coding (why), six types of coding assistance they provide (what), and a five-step workflow in action with practical implementation strategies (how). Additionally, we address the limitations and future outlook of AI in coding. By offering actionable insights, this framework helps to guide researchers in effectively leveraging AI to enhance coding practices and education, accelerating scientific progress.

comment: 23 pages, 7 figure, 3 boxes

♻ ☆ A robust graph-based approach to observational equivalence

Dan R. Ghica, Koko Muroya, Todd Waugh Ambridge

We propose a new step-wise approach to proving observational equivalence, and in particular reasoning about fragility of observational equivalence. Our approach is based on what we call local reasoning. The local reasoning exploits the graphical concept of neighbourhood, and it extracts a new, formal, concept of robustness as a key sufficient condition of observational equivalence. Moreover, our proof methodology is capable of proving a generalised notion of observational equivalence. The generalised notion can be quantified over syntactically restricted contexts instead of all contexts, and also quantitatively constrained in terms of the number of reduction steps. The operational machinery we use is given by a hypergraph-rewriting abstract machine inspired by Girard's Geometry of Interaction. The behaviour of language features, including function abstraction and application, is provided by hypergraph-rewriting rules. We demonstrate our proof methodology using the call-by-value lambda-calculus equipped with (higher-order) state.

Computation and Language 150

☆ Vinoground: Scrutinizing LMMs over Dense Temporal Reasoning with Short Videos

Jianrui Zhang, Mu Cai, Yong Jae Lee

There has been growing sentiment recently that modern large multimodal models (LMMs) have addressed most of the key challenges related to short video comprehension. As a result, both academia and industry are gradually shifting their attention towards the more complex challenges posed by understanding long-form videos. However, is this really the case? Our studies indicate that LMMs still lack many fundamental reasoning capabilities even when dealing with short videos. We introduce Vinoground, a temporal counterfactual LMM evaluation benchmark encompassing 1000 short and natural video-caption pairs. We demonstrate that existing LMMs severely struggle to distinguish temporal differences between different actions and object transformations. For example, the best model GPT-4o only obtains ~50% on our text and video scores, showing a large gap compared to the human baseline of ~90%. All open-source multimodal models and CLIP-based models perform much worse, producing mostly random chance performance. Through this work, we shed light onto the fact that temporal reasoning in short videos is a problem yet to be fully solved. The dataset and evaluation code are available at https://vinoground.github.io.

comment: Project Page: https://vinoground.github.io

☆ Erasing Conceptual Knowledge from Language Models

Rohit Gandikota, Sheridan Feucht, Samuel Marks, David Bau

Concept erasure in language models has traditionally lacked a comprehensive evaluation framework, leading to incomplete assessments of effectiveness of erasure methods. We propose an evaluation paradigm centered on three critical criteria: innocence (complete knowledge removal), seamlessness (maintaining conditional fluent generation), and specificity (preserving unrelated task performance). Our evaluation metrics naturally motivate the development of Erasure of Language Memory (ELM), a new method designed to address all three dimensions. ELM employs targeted low-rank updates to alter output distributions for erased concepts while preserving overall model capabilities including fluency when prompted for an erased concept. We demonstrate ELM's efficacy on biosecurity, cybersecurity, and literary domain erasure tasks. Comparative analysis shows that ELM achieves superior performance across our proposed metrics, including near-random scores on erased topic assessments, generation fluency, maintained accuracy on unrelated benchmarks, and robustness under adversarial attacks. Our code, data, and trained models are available at https://elm.baulab.info

comment: Project Page: https://elm.baulab.info

☆ CorPipe at CRAC 2024: Predicting Zero Mentions from Raw Text

Milan Straka

We present CorPipe 24, the winning entry to the CRAC 2024 Shared Task on Multilingual Coreference Resolution. In this third iteration of the shared task, a novel objective is to also predict empty nodes needed for zero coreference mentions (while the empty nodes were given on input in previous years). This way, coreference resolution can be performed on raw text. We evaluate two model variants: a~two-stage approach (where the empty nodes are predicted first using a pretrained encoder model and then processed together with sentence words by another pretrained model) and a single-stage approach (where a single pretrained encoder model generates empty nodes, coreference mentions, and coreference links jointly). In both settings, CorPipe surpasses other participants by a large margin of 3.9 and 2.8 percent points, respectively. The source code and the trained model are available at https://github.com/ufal/crac2024-corpipe .

comment: Accepted to CRAC 2024

☆ SIEVE: General Purpose Data Filtering System Matching GPT-4o Accuracy at 1% the Cost

Jifan Zhang, Robert Nowak

Creating specialized large language models requires vast amounts of clean, special purpose data for training and fine-tuning. With only a handful of existing large-scale, domain-specific datasets, creation of new datasets is required in most applications. This requires the development of new application-specific filtering of web-scale data. Filtering with a high-performance, general-purpose LLM such as GPT-4o can be highly effective, but this is extremely expensive at web-scale. This paper proposes SIEVE, a lightweight alternative that matches GPT-4o accuracy at a fraction of the cost. SIEVE can perform up to 500 filtering operations for the cost of one GPT-4o filtering call. The key to SIEVE is a seamless integration of GPT-4o and lightweight T5 models, using active learning to fine-tune T5 in the background with a small number of calls to GPT-4o. Once trained, it performs as well as GPT-4o at a tiny fraction of the cost. We experimentally validate SIEVE on the OpenWebText dataset, using five highly customized filter tasks targeting high quality and domain-specific content. Our results demonstrate the effectiveness and efficiency of our method in curating large, high-quality datasets for language model training at a substantially lower cost (1%) than existing techniques. To further validate SIEVE, experiments show that SIEVE and GPT-4o achieve similar accuracy, with human evaluators preferring SIEVE's filtering results to those of GPT-4o.

☆ Training Language Models on Synthetic Edit Sequences Improves Code Synthesis

Ulyana Piterbarg, Lerrel Pinto, Rob Fergus

Software engineers mainly write code by editing existing programs. In contrast, large language models (LLMs) autoregressively synthesize programs in a single pass. One explanation for this is the scarcity of open-sourced edit data. While high-quality instruction data for code synthesis is already scarce, high-quality edit data is even scarcer. To fill this gap, we develop a synthetic data generation algorithm called LintSeq. This algorithm refactors existing code into a sequence of code edits by using a linter to procedurally sample across the error-free insertions that can be used to sequentially write programs. It outputs edit sequences as text strings consisting of consecutive program diffs. To test LintSeq, we use it to refactor a dataset of instruction + program pairs into instruction + program-diff-sequence tuples. Then, we instruction finetune a series of smaller LLMs ranging from 2.6B to 14B parameters on both the re-factored and original versions of this dataset, comparing zero-shot performance on code synthesis benchmarks. We show that during repeated sampling, edit sequence finetuned models produce more diverse programs than baselines. This results in better inference-time scaling for benchmark coverage as a function of samples, i.e. the fraction of problems "pass@k" solved by any attempt given "k" tries. For example, on HumanEval pass@50, small LLMs finetuned on synthetic edit sequences are competitive with GPT-4 and outperform models finetuned on the baseline dataset by +20% (+/-3%) in absolute score. Finally, we also pretrain our own tiny LMs for code understanding. We show that finetuning tiny models on synthetic code edits results in state-of-the-art code synthesis for the on-device model class. Our 150M parameter edit sequence LM matches or outperforms code models with twice as many parameters, both with and without repeated sampling, including Codex and AlphaCode.

☆ CriSPO: Multi-Aspect Critique-Suggestion-guided Automatic Prompt Optimization for Text Generation

Han He, Qianchu Liu, Lei Xu, Chaitanya Shivade, Yi Zhang, Sundararajan Srinivasan, Katrin Kirchhoff

Large language models (LLMs) can generate fluent summaries across domains using prompting techniques, reducing the need to train models for summarization applications. However, crafting effective prompts that guide LLMs to generate summaries with the appropriate level of detail and writing style remains a challenge. In this paper, we explore the use of salient information extracted from the source document to enhance summarization prompts. We show that adding keyphrases in prompts can improve ROUGE F1 and recall, making the generated summaries more similar to the reference and more complete. The number of keyphrases can control the precision-recall trade-off. Furthermore, our analysis reveals that incorporating phrase-level salient information is superior to word- or sentence-level. However, the impact on hallucination is not universally positive across LLMs. To conduct this analysis, we introduce Keyphrase Signal Extractor (CriSPO), a lightweight model that can be finetuned to extract salient keyphrases. By using CriSPO, we achieve consistent ROUGE improvements across datasets and open-weight and proprietary LLMs without any LLM customization. Our findings provide insights into leveraging salient information in building prompt-based summarization systems.

☆ Neutral residues: revisiting adapters for model extension

Franck Signe Talla, Herve Jegou, Edouard Grave

We address the problem of extending a pretrained large language model to a new domain that was not seen at training time, like adding a language for which the original model has seen no or little training data. Popular solutions like fine-tuning or low-rank adaptation are successful at domain adaptation, but formally they do not add any extra capacity and degrade the performance in the original domain. Our paper analyzes this extension problem under three angles: data, architecture and training procedure, which are advantageously considered jointly. In particular, we improve adapters and make it possible to learn an entire new language while ensuring that the output of the neural network is almost unchanged in the original domain. For this purpose, we modify the new residual blocks in a way that leads each new residual block to output near-zeros in the original domain. This solution of neutral residues, which borrows architectural components from mixture of experts, is effective: with only 20% extra learnable weights compared to an original model trained on English, we get results that are significantly better than concurrent approaches (fine-tuning, low-rank or vanilla adapters) in terms of the trade-off between learning a new language and not forgetting English.

☆ MA-RLHF: Reinforcement Learning from Human Feedback with Macro Actions

Yekun Chai, Haoran Sun, Huang Fang, Shuohuan Wang, Yu Sun, Hua Wu

Reinforcement learning from human feedback (RLHF) has demonstrated effectiveness in aligning large language models (LLMs) with human preferences. However, token-level RLHF suffers from the credit assignment problem over long sequences, where delayed rewards make it challenging for the model to discern which actions contributed to successful outcomes. This hinders learning efficiency and slows convergence. In this paper, we propose MA-RLHF, a simple yet effective RLHF framework that incorporates macro actions -- sequences of tokens or higher-level language constructs -- into the learning process. By operating at this higher level of abstraction, our approach reduces the temporal distance between actions and rewards, facilitating faster and more accurate credit assignment. This results in more stable policy gradient estimates and enhances learning efficiency within each episode, all without increasing computational complexity during training or inference. We validate our approach through extensive experiments across various model sizes and tasks, including text summarization, dialogue generation, question answering, and program synthesis. Our method achieves substantial performance improvements over standard RLHF, with performance gains of up to 30% in text summarization and code generation, 18% in dialogue, and 8% in question answering tasks. Notably, our approach reaches parity with vanilla RLHF 1.7x to 2x faster in terms of training time and continues to outperform it with further training. We will make our code and data publicly available at https://github.com/ernie-research/MA-RLHF .

☆ Grounding Large Language Models In Embodied Environment With Imperfect World Models

Haolan Liu, Jishen Zhao

Despite a widespread success in various applications, large language models (LLMs) often stumble when tackling basic physical reasoning or executing robotics tasks, due to a lack of direct experience with the physical nuances of the real world. To address these issues, we propose a Grounding Large language model with Imperfect world MOdel (GLIMO), which utilizes proxy world models such as simulators to collect and synthesize trining data. GLIMO incorporates an LLM agent-based data generator to automatically create high-quality and diverse instruction datasets. The generator includes an iterative self-refining module for temporally consistent experience sampling, a diverse set of question-answering instruction seeds, and a retrieval-augmented generation module for reflecting on prior experiences. Comprehensive experiments show that our approach improve the performance of strong open-source LLMs like LLaMA-3 with a performance boost of 2.04 $\times$, 1.54 $\times$, and 1.82 $\times$ across three different benchmarks, respectively. The performance is able to compete with or surpass their larger counterparts such as GPT-4.

☆ Salient Information Prompting to Steer Content in Prompt-based Abstractive Summarization EMNLP 2024

Lei Xu, Mohammed Asad Karim, Saket Dingliwal, Aparna Elangovan

Large language models (LLMs) can generate fluent summaries across domains using prompting techniques, reducing the need to train models for summarization applications. However, crafting effective prompts that guide LLMs to generate summaries with the appropriate level of detail and writing style remains a challenge. In this paper, we explore the use of salient information extracted from the source document to enhance summarization prompts. We show that adding keyphrases in prompts can improve ROUGE F1 and recall, making the generated summaries more similar to the reference and more complete. The number of keyphrases can control the precision-recall trade-off. Furthermore, our analysis reveals that incorporating phrase-level salient information is superior to word- or sentence-level. However, the impact on hallucination is not universally positive across LLMs. To conduct this analysis, we introduce Keyphrase Signal Extractor (SigExt), a lightweight model that can be finetuned to extract salient keyphrases. By using SigExt, we achieve consistent ROUGE improvements across datasets and open-weight and proprietary LLMs without any LLM customization. Our findings provide insights into leveraging salient information in building prompt-based summarization systems.

comment: Accepted to EMNLP 2024 Industry Track

☆ Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge

Jiayi Ye, Yanbo Wang, Yue Huang, Dongping Chen, Qihui Zhang, Nuno Moniz, Tian Gao, Werner Geyer, Chao Huang, Pin-Yu Chen, Nitesh V Chawla, Xiangliang Zhang

☆ DivScene: Benchmarking LVLMs for Object Navigation with Diverse Scenes and Objects

Zhaowei Wang, Hongming Zhang, Tianqing Fang, Ye Tian, Yue Yang, Kaixin Ma, Xiaoman Pan, Yangqiu Song, Dong Yu

Object navigation in unknown environments is crucial for deploying embodied agents in real-world applications. While we have witnessed huge progress due to large-scale scene datasets, faster simulators, and stronger models, previous studies mainly focus on limited scene types and target objects. In this paper, we study a new task of navigating to diverse target objects in a large number of scene types. To benchmark the problem, we present a large-scale scene dataset, DivScene, which contains 4,614 scenes across 81 different types. With the dataset, we build an end-to-end embodied agent, NatVLM, by fine-tuning a Large Vision Language Model (LVLM) through imitation learning. The LVLM is trained to take previous observations from the environment and generate the next actions. We also introduce CoT explanation traces of the action prediction for better performance when tuning LVLMs. Our extensive experiments find that we can build a performant LVLM-based agent through imitation learning on the shortest paths constructed by a BFS planner without any human supervision. Our agent achieves a success rate that surpasses GPT-4o by over 20%. Meanwhile, we carry out various analyses showing the generalization ability of our agent.

comment: Work in Progress

☆ Unified Multi-Modal Interleaved Document Representation for Information Retrieval

Jaewoo Lee, Joonho Ko, Jinheon Baek, Soyeong Jeong, Sung Ju Hwang

Information Retrieval (IR) methods aim to identify relevant documents in response to a given query, which have gained remarkable attention due to their successful application in various natural language tasks. However, existing approaches typically consider only the textual information within the documents, which overlooks the fact that documents can contain multiple modalities, including texts, images, and tables. Further, they often segment each long document into multiple discrete passages for embedding, preventing them from capturing the overall document context and interactions between paragraphs. We argue that these two limitations lead to suboptimal document representations for retrieval. In this work, to address them, we aim to produce more comprehensive and nuanced document representations by holistically embedding documents interleaved with different modalities. Specifically, we achieve this by leveraging the capability of recent vision-language models that enable the processing and integration of text, images, and tables into a unified format and representation. Moreover, to mitigate the information loss from segmenting documents into passages, instead of representing and retrieving passages individually, we further merge the representations of segmented passages into one single document representation, while we additionally introduce a reranking strategy to decouple and identify the relevant passage within the document if necessary. Then, through extensive experiments on diverse information retrieval scenarios considering both the textual and multimodal queries, we show that our approach substantially outperforms relevant baselines, thanks to the consideration of the multimodal information interleaved within the documents in a unified way.

comment: Preprint

☆ Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation

Rohin Manvi, Anikait Singh, Stefano Ermon

Inference-time computation is a powerful paradigm to enhance the performance of large language models (LLMs), with Best-of-N sampling being a widely used technique. However, this method is computationally expensive, requiring both (1) an external reward model and (2) the generation of multiple samples. In this work, we introduce a new generative self-evaluation scheme designed to adaptively reduce the number of generated samples while maintaining or even improving performance. We use a generative reward model formulation, allowing the LLM to predict mid-generation the probability that restarting the generation will yield a better response. These predictions are obtained without an external reward model and can be used to decide whether or not to generate more samples, prune unpromising samples early on, or to pick the best sample. This capability is very inexpensive as it involves generating a single predefined token. Trained using a dataset constructed with real unfiltered LMSYS user prompts, Llama 3.1 8B's win rate against GPT-4 on AlpacaEval increases from 21% to 34% with 16 samples and math performance on GSM8K improves from 84% to 91%. By sampling only when the LLM determines that it is beneficial to do so and adaptively adjusting temperature annealing, we demonstrate that 74% of the improvement from using 16 samples can be achieved with only 1.2 samples on average. We further demonstrate that 50-75% of samples can be pruned early in generation with minimal degradation in performance. Overall, our methods enable more efficient and scalable compute utilization during inference for LLMs.

☆ Large Language Models as Markov Chains

Oussama Zekri, Ambroise Odonnat, Abdelhakim Benechehab, Linus Bleistein, Nicolas Boullé, Ievgen Redko

Large language models (LLMs) have proven to be remarkably efficient, both across a wide range of natural language processing tasks and well beyond them. However, a comprehensive theoretical analysis of the origins of their impressive performance remains elusive. In this paper, we approach this challenging task by drawing an equivalence between generic autoregressive language models with vocabulary of size $T$ and context window of size $K$ and Markov chains defined on a finite state space of size $\mathcal{O}(T^K)$. We derive several surprising findings related to the existence of a stationary distribution of Markov chains that capture the inference power of LLMs, their speed of convergence to it, and the influence of the temperature on the latter. We then prove pre-training and in-context generalization bounds and show how the drawn equivalence allows us to enrich their interpretation. Finally, we illustrate our theoretical guarantees with experiments on several recent LLMs to highlight how they capture the behavior observed in practice.

comment: 49 pages, 17 figures

☆ Domain-Specific Retrieval-Augmented Generation Using Vector Stores, Knowledge Graphs, and Tensor Factorization ICML

Ryan C. Barron, Ves Grantcharov, Selma Wanna, Maksim E. Eren, Manish Bhattarai, Nicholas Solovyev, George Tompkins, Charles Nicholas, Kim Ø. Rasmussen, Cynthia Matuszek, Boian S. Alexandrov

Large Language Models (LLMs) are pre-trained on large-scale corpora and excel in numerous general natural language processing (NLP) tasks, such as question answering (QA). Despite their advanced language capabilities, when it comes to domain-specific and knowledge-intensive tasks, LLMs suffer from hallucinations, knowledge cut-offs, and lack of knowledge attributions. Additionally, fine tuning LLMs' intrinsic knowledge to highly specific domains is an expensive and time consuming process. The retrieval-augmented generation (RAG) process has recently emerged as a method capable of optimization of LLM responses, by referencing them to a predetermined ontology. It was shown that using a Knowledge Graph (KG) ontology for RAG improves the QA accuracy, by taking into account relevant sub-graphs that preserve the information in a structured manner. In this paper, we introduce SMART-SLIC, a highly domain-specific LLM framework, that integrates RAG with KG and a vector store (VS) that store factual domain specific information. Importantly, to avoid hallucinations in the KG, we build these highly domain-specific KGs and VSs without the use of LLMs, but via NLP, data mining, and nonnegative tensor factorization with automatic model selection. Pairing our RAG with a domain-specific: (i) KG (containing structured information), and (ii) VS (containing unstructured information) enables the development of domain-specific chat-bots that attribute the source of information, mitigate hallucinations, lessen the need for fine-tuning, and excel in highly domain-specific question answering tasks. We pair SMART-SLIC with chain-of-thought prompting agents. The framework is designed to be generalizable to adapt to any specific or specialized domain. In this paper, we demonstrate the question answering capabilities of our framework on a corpus of scientific publications on malware analysis and anomaly detection.

comment: 9 pages 7 figures, 1 table, 1 cypher code Accepted to ICMLA 2024

☆ UncertaintyRAG: Span-Level Uncertainty Enhanced Long-Context Modeling for Retrieval-Augmented Generation

Zixuan Li, Jing Xiong, Fanghua Ye, Chuanyang Zheng, Xun Wu, Jianqiao Lu, Zhongwei Wan, Xiaodan Liang, Chengming Li, Zhenan Sun, Lingpeng Kong, Ngai Wong

We present UncertaintyRAG, a novel approach for long-context Retrieval-Augmented Generation (RAG) that utilizes Signal-to-Noise Ratio (SNR)-based span uncertainty to estimate similarity between text chunks. This span uncertainty enhances model calibration, improving robustness and mitigating semantic inconsistencies introduced by random chunking. Leveraging this insight, we propose an efficient unsupervised learning technique to train the retrieval model, alongside an effective data sampling and scaling strategy. UncertaintyRAG outperforms baselines by 2.03% on LLaMA-2-7B, achieving state-of-the-art results while using only 4% of the training data compared to other advanced open-source retrieval models under distribution shift settings. Our method demonstrates strong calibration through span uncertainty, leading to improved generalization and robustness in long-context RAG tasks. Additionally, UncertaintyRAG provides a lightweight retrieval model that can be integrated into any large language model with varying context window lengths, without the need for fine-tuning, showcasing the flexibility of our approach.

☆ Video Instruction Tuning With Synthetic Data

Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, Chunyuan Li

comment: Project page: https://llava-vl.github.io/blog/2024-09-30-llava-video/

☆ LLaVA-Critic: Learning to Evaluate Multimodal Models

Tianyi Xiong, Xiyao Wang, Dong Guo, Qinghao Ye, Haoqi Fan, Quanquan Gu, Heng Huang, Chunyuan Li

We introduce LLaVA-Critic, the first open-source large multimodal model (LMM) designed as a generalist evaluator to assess performance across a wide range of multimodal tasks. LLaVA-Critic is trained using a high-quality critic instruction-following dataset that incorporates diverse evaluation criteria and scenarios. Our experiments demonstrate the model's effectiveness in two key areas: (1) LMM-as-a-Judge, where LLaVA-Critic provides reliable evaluation scores, performing on par with or surpassing GPT models on multiple evaluation benchmarks; and (2) Preference Learning, where it generates reward signals for preference learning, enhancing model alignment capabilities. This work underscores the potential of open-source LMMs in self-critique and evaluation, setting the stage for future research into scalable, superhuman alignment feedback mechanisms for LMMs.

comment: Project Page: https://llava-vl.github.io/blog/2024-10-03-llava-critic

☆ LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations

Hadas Orgad, Michael Toker, Zorik Gekhman, Roi Reichart, Idan Szpektor, Hadas Kotek, Yonatan Belinkov

☆ Selective Attention Improves Transformer

Yaniv Leviathan, Matan Kalman, Yossi Matias

Unneeded elements in the attention's context degrade performance. We introduce Selective Attention, a simple parameter-free change to the standard attention mechanism which reduces attention to unneeded elements. Selective attention improves language modeling performance in a variety of model sizes and context lengths. For example, a range of transformers trained with the language modeling objective on C4 with selective attention perform equivalently to standard transformers with ~2X more heads and parameters in their attention modules. Selective attention also allows decreasing the size of the attention's context buffer, leading to meaningful reductions in the memory and compute requirements during inference. For example, transformers with 100M parameters trained on C4 with context sizes of 512, 1,024, and 2,048 need 16X, 25X, and 47X less memory for their attention module, respectively, when equipped with selective attention, as those without selective attention, with the same validation perplexity.

☆ HELMET: How to Evaluate Long-Context Language Models Effectively and Thoroughly

Howard Yen, Tianyu Gao, Minmin Hou, Ke Ding, Daniel Fleischer, Peter Izasak, Moshe Wasserblat, Danqi Chen

There have been many benchmarks for evaluating long-context language models (LCLMs), but developers often rely on synthetic tasks like needle-in-a-haystack (NIAH) or arbitrary subsets of tasks. It remains unclear whether they translate to the diverse downstream applications of LCLMs, and the inconsistency further complicates model comparison. We investigate the underlying reasons behind current practices and find that existing benchmarks often provide noisy signals due to low coverage of applications, insufficient lengths, unreliable metrics, and incompatibility with base models. In this work, we present HELMET (How to Evaluate Long-context Models Effectively and Thoroughly), a comprehensive benchmark encompassing seven diverse, application-centric categories. We also address many issues in previous benchmarks by adding controllable lengths up to 128k tokens, model-based evaluation for reliable metrics, and few-shot prompting for robustly evaluating base models. Consequently, we demonstrate that HELMET offers more reliable and consistent rankings of frontier LCLMs. Through a comprehensive study of 51 LCLMs, we find that (1) synthetic tasks like NIAH are not good predictors of downstream performance; (2) the diverse categories in HELMET exhibit distinct trends and low correlation with each other; and (3) while most LCLMs achieve perfect NIAH scores, open-source models significantly lag behind closed ones when the task requires full-context reasoning or following complex instructions -- the gap widens with increased lengths. Finally, we recommend using our RAG tasks for fast model development, as they are easy to run and more predictive of other downstream performance; ultimately, we advocate for a holistic evaluation across diverse tasks.

comment: Code and data are available here: https://github.com/princeton-nlp/HELMET

☆ On the Proper Treatment of Tokenization in Psycholinguistics EMNLP 2024

Mario Giulianelli, Luca Malagutti, Juan Luis Gastaldi, Brian DuSell, Tim Vieira, Ryan Cotterell

Language models are widely used in computational psycholinguistics to test theories that relate the negative log probability (the surprisal) of a region of interest (a substring of characters) under a language model to its cognitive cost experienced by readers, as operationalized, for example, by gaze duration on the region. However, the application of modern language models to psycholinguistic studies is complicated by the practice of using tokenization as an intermediate step in training a model. Doing so results in a language model over token strings rather than one over character strings. Vexingly, regions of interest are generally misaligned with these token strings. The paper argues that token-level language models should be (approximately) marginalized into character-level language models before they are used in psycholinguistic studies to compute the surprisal of a region of interest; then, the marginalized character-level language model can be used to compute the surprisal of an arbitrary character substring, which we term a focal area, that the experimenter may wish to use as a predictor. Our proposal of marginalizing a token-level model into a character-level one solves this misalignment issue independently of the tokenization scheme. Empirically, we discover various focal areas whose surprisal is a better psychometric predictor than the surprisal of the region of interest itself.

comment: Main conference long paper at EMNLP 2024

☆ HiddenGuard: Fine-Grained Safe Generation with Specialized Representation Router

Lingrui Mei, Shenghua Liu, Yiwei Wang, Baolong Bi, Ruibin Yuan, Xueqi Cheng

As Large Language Models (LLMs) grow increasingly powerful, ensuring their safety and alignment with human values remains a critical challenge. Ideally, LLMs should provide informative responses while avoiding the disclosure of harmful or sensitive information. However, current alignment approaches, which rely heavily on refusal strategies, such as training models to completely reject harmful prompts or applying coarse filters are limited by their binary nature. These methods either fully deny access to information or grant it without sufficient nuance, leading to overly cautious responses or failures to detect subtle harmful content. For example, LLMs may refuse to provide basic, public information about medication due to misuse concerns. Moreover, these refusal-based methods struggle to handle mixed-content scenarios and lack the ability to adapt to context-dependent sensitivities, which can result in over-censorship of benign content. To overcome these challenges, we introduce HiddenGuard, a novel framework for fine-grained, safe generation in LLMs. HiddenGuard incorporates Prism (rePresentation Router for In-Stream Moderation), which operates alongside the LLM to enable real-time, token-level detection and redaction of harmful content by leveraging intermediate hidden states. This fine-grained approach allows for more nuanced, context-aware moderation, enabling the model to generate informative responses while selectively redacting or replacing sensitive information, rather than outright refusal. We also contribute a comprehensive dataset with token-level fine-grained annotations of potentially harmful information across diverse contexts. Our experiments demonstrate that HiddenGuard achieves over 90% in F1 score for detecting and redacting harmful content while preserving the overall utility and informativeness of the model's responses.

☆ DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Life

Yu Ying Chiu, Liwei Jiang, Yejin Choi

As we increasingly seek guidance from LLMs for decision-making in daily life, many of these decisions are not clear-cut and depend significantly on the personal values and ethical standards of the users. We present DailyDilemmas, a dataset of 1,360 moral dilemmas encountered in everyday life. Each dilemma includes two possible actions and with each action, the affected parties and human values invoked. Based on these dilemmas, we consolidated a set of human values across everyday topics e.g., interpersonal relationships, workplace, and environmental issues. We evaluated LLMs on these dilemmas to determine what action they will take and the values represented by these actions. Then, we analyzed these values through the lens of five popular theories inspired by sociology, psychology and philosophy. These theories are: World Value Survey, Moral Foundation Theory, Maslow's Hierarchy of Needs, Aristotle's Virtues, and Plutchik Wheel of Emotion. We find that LLMs are most aligned with the self-expression over survival values in terms of World Value Survey, care over loyalty in Moral Foundation Theory. Interestingly, we find large preferences differences in models for some core values such as truthfulness e.g., Mixtral-8x7B model tends to neglect it by 9.7% while GPT-4-turbo model tends to select it by 9.4%. We also study the recent guidance released by OpenAI (ModelSpec), and Anthropic (Constitutional AI) to understand how their released principles reflect their actual value prioritization when facing nuanced moral reasoning in daily-life settings. We find that end users cannot effectively steer such prioritization using system prompts.

comment: Preprint. Under Review

☆ Distilling an End-to-End Voice Assistant Without Instruction Training Data

William Held, Ella Li, Michael Ryan, Weiyan Shi, Yanzhe Zhang, Diyi Yang

Voice assistants, such as Siri and Google Assistant, typically model audio and text separately, resulting in lost speech information and increased complexity. Recent efforts to address this with end-to-end Speech Large Language Models (LLMs) trained with supervised finetuning (SFT) have led to models ``forgetting" capabilities from text-only LLMs. Our work proposes an alternative paradigm for training Speech LLMs without instruction data, using the response of a text-only LLM to transcripts as self-supervision. Importantly, this process can be performed without annotated responses. We show that our Distilled Voice Assistant (DiVA) generalizes to Spoken Question Answering, Classification, and Translation. Furthermore, we show that DiVA better meets user preferences, achieving a 72\% win rate compared with state-of-the-art models like Qwen 2 Audio, despite using $>$100x less training compute.

☆ CulturalBench: a Robust, Diverse and Challenging Benchmark on Measuring the (Lack of) Cultural Knowledge of LLMs

Yu Ying Chiu, Liwei Jiang, Bill Yuchen Lin, Chan Young Park, Shuyue Stella Li, Sahithya Ravi, Mehar Bhatia, Maria Antoniak, Yulia Tsvetkov, Vered Shwartz, Yejin Choi

To make large language models (LLMs) more helpful across diverse cultures, it is essential to have effective cultural knowledge benchmarks to measure and track our progress. Effective benchmarks need to be robust, diverse, and challenging. We introduce CulturalBench: a set of 1,227 human-written and human-verified questions for effectively assessing LLMs' cultural knowledge, covering 45 global regions including the underrepresented ones like Bangladesh, Zimbabwe, and Peru. Questions - each verified by five independent annotators - span 17 diverse topics ranging from food preferences to greeting etiquettes. We evaluate models on two setups: CulturalBench-Easy and CulturalBench-Hard which share the same questions but asked differently. We find that LLMs are sensitive to such difference in setups (e.g., GPT-4o with 27.3% difference). Compared to human performance (92.6% accuracy), CulturalBench-Hard is more challenging for frontier LLMs with the best performing model (GPT-4o) at only 61.5% and the worst (Llama3-8b) at 21.4%. Moreover, we find that LLMs often struggle with tricky questions that have multiple correct answers (e.g., What utensils do the Chinese usually use?), revealing a tendency to converge to a single answer. Our results also indicate that OpenAI GPT-4o substantially outperform other proprietary and open source models in questions related to all but one region (Oceania). Nonetheless, all models consistently underperform on questions related to South America and the Middle East.

comment: Preprint. Under review

☆ FAN: Fourier Analysis Networks

Yihong Dong, Ge Li, Yongding Tao, Xue Jiang, Kechi Zhang, Jia Li, Jing Su, Jun Zhang, Jingjing Xu

Despite the remarkable success achieved by neural networks, particularly those represented by MLP and Transformer, we reveal that they exhibit potential flaws in the modeling and reasoning of periodicity, i.e., they tend to memorize the periodic data rather than genuinely understanding the underlying principles of periodicity. However, periodicity is a crucial trait in various forms of reasoning and generalization, underpinning predictability across natural and engineered systems through recurring patterns in observations. In this paper, we propose FAN, a novel network architecture based on Fourier Analysis, which empowers the ability to efficiently model and reason about periodic phenomena. By introducing Fourier Series, the periodicity is naturally integrated into the structure and computational processes of the neural network, thus achieving a more accurate expression and prediction of periodic patterns. As a promising substitute to multi-layer perceptron (MLP), FAN can seamlessly replace MLP in various models with fewer parameters and FLOPs. Through extensive experiments, we demonstrate the effectiveness of FAN in modeling and reasoning about periodic functions, and the superiority and generalizability of FAN across a range of real-world tasks, including symbolic formula representation, time series forecasting, and language modeling.

☆ Examining Language Modeling Assumptions Using an Annotated Literary Dialect Corpus EMNLP2024

Craig Messner, Tom Lippincott

We present a dataset of 19th century American literary orthovariant tokens with a novel layer of human-annotated dialect group tags designed to serve as the basis for computational experiments exploring literarily meaningful orthographic variation. We perform an initial broad set of experiments over this dataset using both token (BERT) and character (CANINE)-level contextual language models. We find indications that the "dialect effect" produced by intentional orthographic variation employs multiple linguistic channels, and that these channels are able to be surfaced to varied degrees given particular language modelling assumptions. Specifically, we find evidence showing that choice of tokenization scheme meaningfully impact the type of orthographic information a model is able to surface.

comment: Accepted to NLP4DH@EMNLP2024

☆ How to Train Long-Context Language Models (Effectively)

Tianyu Gao, Alexander Wettig, Howard Yen, Danqi Chen

We study continued training and supervised fine-tuning (SFT) of a language model (LM) to make effective use of long-context information. We first establish a reliable evaluation protocol to guide model development -- Instead of perplexity or simple needle-in-a-haystack (NIAH) tests, we use a broad set of long-context tasks, and we evaluate models after SFT with instruction data as this better reveals long-context abilities. Supported by our robust evaluations, we run thorough experiments to decide the data mix for continued pre-training, the instruction tuning dataset, and many other design choices. We find that (1) code repositories and books are excellent sources of long data, but it is crucial to combine them with high-quality short data; (2) training with a sequence length beyond the evaluation length boosts long-context performance; (3) for SFT, using only short instruction datasets yields strong performance on long-context tasks. Our final model, ProLong-8B, which is initialized from Llama-3 and trained on 40B tokens, demonstrates state-of-the-art long-context performance among similarly sized models at a length of 128K. ProLong outperforms Llama-3.18B-Instruct on the majority of long-context tasks despite having seen only 5% as many tokens during long-context training. Additionally, ProLong can effectively process up to 512K tokens, one of the longest context windows of publicly available LMs.

comment: Our code, data, and models are available at https://github.com/princeton-nlp/ProLong

☆ Hate Personified: Investigating the role of LLMs in content moderation EMNLP'24

Sarah Masud, Sahajpreet Singh, Viktor Hangya, Alexander Fraser, Tanmoy Chakraborty

For subjective tasks such as hate detection, where people perceive hate differently, the Large Language Model's (LLM) ability to represent diverse groups is unclear. By including additional context in prompts, we comprehensively analyze LLM's sensitivity to geographical priming, persona attributes, and numerical information to assess how well the needs of various groups are reflected. Our findings on two LLMs, five languages, and six datasets reveal that mimicking persona-based attributes leads to annotation variability. Meanwhile, incorporating geographical signals leads to better regional alignment. We also find that the LLMs are sensitive to numerical anchors, indicating the ability to leverage community-based flagging efforts and exposure to adversaries. Our work provides preliminary guidelines and highlights the nuances of applying LLMs in culturally sensitive cases.

comment: 17 pages, 6 Figures, 13 Tables, EMNLP'24 Mains

☆ Measuring and Improving Persuasiveness of Generative Models

Somesh Singh, Yaman K Singla, Harini SI, Balaji Krishnamurthy

LLMs are increasingly being used in workflows involving generating content to be consumed by humans (e.g., marketing) and also in directly interacting with humans (e.g., through chatbots). The development of such systems that are capable of generating verifiably persuasive messages presents both opportunities and challenges for society. On the one hand, such systems could positively impact domains like advertising and social good, such as addressing drug addiction, and on the other, they could be misused for spreading misinformation and shaping political opinions. To channel LLMs' impact on society, we need to develop systems to measure and benchmark their persuasiveness. With this motivation, we introduce PersuasionBench and PersuasionArena, the first large-scale benchmark and arena containing a battery of tasks to measure the persuasion ability of generative models automatically. We investigate to what extent LLMs know and leverage linguistic patterns that can help them generate more persuasive language. Our findings indicate that the persuasiveness of LLMs correlates positively with model size, but smaller models can also be made to have a higher persuasiveness than much larger models. Notably, targeted training using synthetic and natural datasets significantly enhances smaller models' persuasive capabilities, challenging scale-dependent assumptions. Our findings carry key implications for both model developers and policymakers. For instance, while the EU AI Act and California's SB-1047 aim to regulate AI models based on the number of floating point operations, we demonstrate that simple metrics like this alone fail to capture the full scope of AI's societal impact. We invite the community to explore and contribute to PersuasionArena and PersuasionBench, available at https://bit.ly/measure-persuasion, to advance our understanding of AI-driven persuasion and its societal implications.

☆ Undesirable Memorization in Large Language Models: A Survey

Ali Satvaty, Suzan Verberne, Fatih Turkmen

While recent research increasingly showcases the remarkable capabilities of Large Language Models (LLMs), it's vital to confront their hidden pitfalls. Among these challenges, the issue of memorization stands out, posing significant ethical and legal risks. In this paper, we presents a Systematization of Knowledge (SoK) on the topic of memorization in LLMs. Memorization is the effect that a model tends to store and reproduce phrases or passages from the training data and has been shown to be the fundamental issue to various privacy and security attacks against LLMs. We begin by providing an overview of the literature on the memorization, exploring it across five key dimensions: intentionality, degree, retrievability, abstraction, and transparency. Next, we discuss the metrics and methods used to measure memorization, followed by an analysis of the factors that contribute to memorization phenomenon. We then examine how memorization manifests itself in specific model architectures and explore strategies for mitigating these effects. We conclude our overview by identifying potential research topics for the near future: to develop methods for balancing performance and privacy in LLMs, and the analysis of memorization in specific contexts, including conversational agents, retrieval-augmented generation, multilingual language models, and diffusion language models.

☆ Immunogenicity Prediction with Dual Attention Enables Vaccine Target Selection

Song Li, Yang Tan, Song Ke, Liang Hong, Bingxin Zhou

Immunogenicity prediction is a central topic in reverse vaccinology for finding candidate vaccines that can trigger protective immune responses. Existing approaches typically rely on highly compressed features and simple model architectures, leading to limited prediction accuracy and poor generalizability. To address these challenges, we introduce ProVaccine, a novel deep learning solution with a dual attention mechanism that integrates pre-trained latent vector representations of protein sequences and structures. We also compile the most comprehensive immunogenicity dataset to date, encompassing over 9,500 antigen sequences, structures, and immunogenicity labels from bacteria, viruses, and tumors. Extensive experiments demonstrate that ProVaccine outperforms existing methods across a wide range of evaluation metrics. Furthermore, we establish a post-hoc validation protocol to assess the practical significance of deep learning models in tackling vaccine design challenges. Our work provides an effective tool for vaccine design and sets valuable benchmarks for future research.

comment: 18 pages, 11 tables, 5 figures

☆ Attention in Large Language Models Yields Efficient Zero-Shot Re-Rankers

Shijie Chen, Bernal Jiménez Gutiérrez, Yu Su

Information retrieval (IR) systems have played a vital role in modern digital life and have cemented their continued usefulness in this new era of generative AI via retrieval-augmented generation. With strong language processing capabilities and remarkable versatility, large language models (LLMs) have become popular choices for zero-shot re-ranking in IR systems. So far, LLM-based re-ranking methods rely on strong generative capabilities, which restricts their use to either specialized or powerful proprietary models. Given these restrictions, we ask: is autoregressive generation necessary and optimal for LLMs to perform re-ranking? We hypothesize that there are abundant signals relevant to re-ranking within LLMs that might not be used to their full potential via generation. To more directly leverage such signals, we propose in-context re-ranking (ICR), a novel method that leverages the change in attention pattern caused by the search query for accurate and efficient re-ranking. To mitigate the intrinsic biases in LLMs, we propose a calibration method using a content-free query. Due to the absence of generation, ICR only requires two ($O(1)$) forward passes to re-rank $N$ documents, making it substantially more efficient than generative re-ranking methods that require at least $O(N)$ forward passes. Our novel design also enables ICR to be applied to any LLM without specialized training while guaranteeing a well-formed ranking. Extensive experiments with two popular open-weight LLMs on standard single-hop and multi-hop information retrieval benchmarks show that ICR outperforms RankGPT while cutting the latency by more than 60% in practice. Through detailed analyses, we show that ICR's performance is specially strong on tasks that require more complex re-ranking signals. Our findings call for further exploration on novel ways of utilizing open-weight LLMs beyond text generation.

☆ Large Language Model for Multi-Domain Translation: Benchmarking and Domain CoT Fine-tuning

Tianxiang Hu, Pei Zhang, Baosong Yang, Jun Xie, Derek F. Wong, Rui Wang

Achieving consistent high-quality machine translation (MT) across diverse domains remains a significant challenge, primarily due to the limited and imbalanced parallel training data available in various domains. While large language models (LLMs) have demonstrated impressive general understanding and generation abilities, their potential in multi-domain MT is under-explored. We establish a comprehensive benchmark for multi-domain translation, featuring 25 German$\Leftrightarrow$English and 22 Chinese$\Leftrightarrow$English test sets respectively covering 15 domains. Our evaluation of prominent LLMs reveals a discernible performance gap against traditional MT systems, highlighting domain overfitting and catastrophic forgetting issues after fine-tuning on domain-limited corpora. To mitigate this, we propose a domain Chain of Thought (CoT) fine-tuning technique that utilizes the intrinsic multi-domain intelligence of LLMs to improve translation performance. This method inspires the LLM to perceive domain information from the source text, which then serves as a helpful hint to guide the translation process. Despite being trained on a small dataset of four domains, our CoT fine-tune approach achieves notable enhancements in translation accuracy and domain robustness than traditional fine-tuning, as evidenced by an average 1.53 BLEU score increase in over 20 German$\rightarrow$English distinct out-of-domain tests.

☆ NL-Eye: Abductive NLI for Images

Mor Ventura, Michael Toker, Nitay Calderon, Zorik Gekhman, Yonatan Bitton, Roi Reichart

Will a Visual Language Model (VLM)-based bot warn us about slipping if it detects a wet floor? Recent VLMs have demonstrated impressive capabilities, yet their ability to infer outcomes and causes remains underexplored. To address this, we introduce NL-Eye, a benchmark designed to assess VLMs' visual abductive reasoning skills. NL-Eye adapts the abductive Natural Language Inference (NLI) task to the visual domain, requiring models to evaluate the plausibility of hypothesis images based on a premise image and explain their decisions. NL-Eye consists of 350 carefully curated triplet examples (1,050 images) spanning diverse reasoning categories: physical, functional, logical, emotional, cultural, and social. The data curation process involved two steps - writing textual descriptions and generating images using text-to-image models, both requiring substantial human involvement to ensure high-quality and challenging scenes. Our experiments show that VLMs struggle significantly on NL-Eye, often performing at random baseline levels, while humans excel in both plausibility prediction and explanation quality. This demonstrates a deficiency in the abductive reasoning capabilities of modern VLMs. NL-Eye represents a crucial step toward developing VLMs capable of robust multimodal reasoning for real-world applications, including accident-prevention bots and generated video verification.

☆ IndicSentEval: How Effectively do Multilingual Transformer Models encode Linguistic Properties for Indic Languages?

Akhilesh Aravapalli, Mounika Marreddy, Subba Reddy Oota, Radhika Mamidi, Manish Gupta

Transformer-based models have revolutionized the field of natural language processing. To understand why they perform so well and to assess their reliability, several studies have focused on questions such as: Which linguistic properties are encoded by these models, and to what extent? How robust are these models in encoding linguistic properties when faced with perturbations in the input text? However, these studies have mainly focused on BERT and the English language. In this paper, we investigate similar questions regarding encoding capability and robustness for 8 linguistic properties across 13 different perturbations in 6 Indic languages, using 9 multilingual Transformer models (7 universal and 2 Indic-specific). To conduct this study, we introduce a novel multilingual benchmark dataset, IndicSentEval, containing approximately $\sim$47K sentences. Surprisingly, our probing analysis of surface, syntactic, and semantic properties reveals that while almost all multilingual models demonstrate consistent encoding performance for English, they show mixed results for Indic languages. As expected, Indic-specific multilingual models capture linguistic properties in Indic languages better than universal models. Intriguingly, universal models broadly exhibit better robustness compared to Indic-specific models, particularly under perturbations such as dropping both nouns and verbs, dropping only verbs, or keeping only nouns. Overall, this study provides valuable insights into probing and perturbation-specific strengths and weaknesses of popular multilingual Transformer-based models for different Indic languages. We make our code and dataset publicly available [https://tinyurl.com/IndicSentEval}].

comment: 23 pages, 11 figures

☆ Ethio-Fake: Cutting-Edge Approaches to Combat Fake News in Under-Resourced Languages Using Explainable AI

Mesay Gemeda Yigezu, Melkamu Abay Mersha, Girma Yohannis Bade, Jugal Kalita, Olga Kolesnikova, Alexander Gelbukh

The proliferation of fake news has emerged as a significant threat to the integrity of information dissemination, particularly on social media platforms. Misinformation can spread quickly due to the ease of creating and disseminating content, affecting public opinion and sociopolitical events. Identifying false information is therefore essential to reducing its negative consequences and maintaining the reliability of online news sources. Traditional approaches to fake news detection often rely solely on content-based features, overlooking the crucial role of social context in shaping the perception and propagation of news articles. In this paper, we propose a comprehensive approach that integrates social context-based features with news content features to enhance the accuracy of fake news detection in under-resourced languages. We perform several experiments utilizing a variety of methodologies, including traditional machine learning, neural networks, ensemble learning, and transfer learning. Assessment of the outcomes of the experiments shows that the ensemble learning approach has the highest accuracy, achieving a 0.99 F1 score. Additionally, when compared with monolingual models, the fine-tuned model with the target language outperformed others, achieving a 0.94 F1 score. We analyze the functioning of the models, considering the important features that contribute to model performance, using explainable AI techniques.

☆ Agents' Room: Narrative Generation through Multi-step Collaboration ICLR 2025

Fantine Huot, Reinald Kim Amplayo, Jennimaria Palomaki, Alice Shoshana Jakobovits, Elizabeth Clark, Mirella Lapata

Writing compelling fiction is a multifaceted process combining elements such as crafting a plot, developing interesting characters, and using evocative language. While large language models (LLMs) show promise for story writing, they currently rely heavily on intricate prompting, which limits their use. We propose Agents' Room, a generation framework inspired by narrative theory, that decomposes narrative writing into subtasks tackled by specialized agents. To illustrate our method, we introduce Tell Me A Story, a high-quality dataset of complex writing prompts and human-written stories, and a novel evaluation framework designed specifically for assessing long narratives. We show that Agents' Room generates stories that are preferred by expert evaluators over those produced by baseline systems by leveraging collaboration and specialization to decompose the complex story writing task into tractable components. We provide extensive analysis with automated and human-based metrics of the generated output.

comment: Under review as a conference paper at ICLR 2025

☆ Towards Implicit Bias Detection and Mitigation in Multi-Agent LLM Interactions EMNLP

Angana Borah, Rada Mihalcea

As Large Language Models (LLMs) continue to evolve, they are increasingly being employed in numerous studies to simulate societies and execute diverse social tasks. However, LLMs are susceptible to societal biases due to their exposure to human-generated data. Given that LLMs are being used to gain insights into various societal aspects, it is essential to mitigate these biases. To that end, our study investigates the presence of implicit gender biases in multi-agent LLM interactions and proposes two strategies to mitigate these biases. We begin by creating a dataset of scenarios where implicit gender biases might arise, and subsequently develop a metric to assess the presence of biases. Our empirical analysis reveals that LLMs generate outputs characterized by strong implicit bias associations (>= 50\% of the time). Furthermore, these biases tend to escalate following multi-agent interactions. To mitigate them, we propose two strategies: self-reflection with in-context examples (ICE); and supervised fine-tuning. Our research demonstrates that both methods effectively mitigate implicit biases, with the ensemble of fine-tuning and self-reflection proving to be the most successful.

comment: Accepted to EMNLP Findings 2024

☆ Convolutional Variational Autoencoders for Spectrogram Compression in Automatic Speech Recognition

Olga Yakovenko, Ivan Bondarenko

comment: Theory and Practice of Natural Computing 9th International Conference, TPNC 2020, Taoyuan, Taiwan, 2020, Proceedings 9

☆ Improving Unsupervised Constituency Parsing via Maximizing Semantic Information

Junjie Chen, Xiangheng He, Yusuke Miyao, Danushka Bollegala

Unsupervised constituency parsers organize phrases within a sentence into a tree-shaped syntactic constituent structure that reflects the organization of sentence semantics. However, the traditional objective of maximizing sentence log-likelihood (LL) does not explicitly account for the close relationship between the constituent structure and the semantics, resulting in a weak correlation between LL values and parsing accuracy. In this paper, we introduce a novel objective for training unsupervised parsers: maximizing the information between constituent structures and sentence semantics (SemInfo). We introduce a bag-of-substrings model to represent the semantics and apply the probability-weighted information metric to estimate the SemInfo. Additionally, we develop a Tree Conditional Random Field (TreeCRF)-based model to apply the SemInfo maximization objective to Probabilistic Context-Free Grammar (PCFG) induction, the state-of-the-art method for unsupervised constituency parsing. Experiments demonstrate that SemInfo correlates more strongly with parsing accuracy than LL. Our algorithm significantly enhances parsing accuracy by an average of 7.85 points across five PCFG variants and in four languages, achieving new state-of-the-art results in three of the four languages.

☆ ColaCare: Enhancing Electronic Health Record Modeling through Large Language Model-Driven Multi-Agent Collaboration

Zixiang Wang, Yinghao Zhu, Huiya Zhao, Xiaochen Zheng, Tianlong Wang, Wen Tang, Yasha Wang, Chengwei Pan, Ewen M. Harrison, Junyi Gao, Liantao Ma

We introduce ColaCare, a framework that enhances Electronic Health Record (EHR) modeling through multi-agent collaboration driven by Large Language Models (LLMs). Our approach seamlessly integrates domain-specific expert models with LLMs to bridge the gap between structured EHR data and text-based reasoning. Inspired by clinical consultations, ColaCare employs two types of agents: DoctorAgent and MetaAgent, which collaboratively analyze patient data. Expert models process and generate predictions from numerical EHR data, while LLM agents produce reasoning references and decision-making reports within the collaborative consultation framework. We additionally incorporate the Merck Manual of Diagnosis and Therapy (MSD) medical guideline within a retrieval-augmented generation (RAG) module for authoritative evidence support. Extensive experiments conducted on four distinct EHR datasets demonstrate ColaCare's superior performance in mortality prediction tasks, underscoring its potential to revolutionize clinical decision support systems and advance personalized precision medicine. The code, complete prompt templates, more case studies, etc. are publicly available at the anonymous link: https://colacare.netlify.app.

☆ MedVisionLlama: Leveraging Pre-Trained Large Language Model Layers to Enhance Medical Image Segmentation WACV

Gurucharan Marthi Krishna Kumar, Aman Chadha, Janine Mendola, Amir Shmuel

Large Language Models (LLMs), known for their versatility in textual data, are increasingly being explored for their potential to enhance medical image segmentation, a crucial task for accurate diagnostic imaging. This study explores enhancing Vision Transformers (ViTs) for medical image segmentation by integrating pre-trained LLM transformer blocks. Our approach, which incorporates a frozen LLM transformer block into the encoder of a ViT-based model, leads to substantial improvements in segmentation performance across various medical imaging modalities. We propose a Hybrid Attention Mechanism that combines global and local feature learning with a Multi-Scale Fusion Block for aggregating features across different scales. The enhanced model shows significant performance gains, including an average Dice score increase from 0.74 to 0.79 and improvements in accuracy, precision, and the Jaccard Index. These results demonstrate the effectiveness of LLM-based transformers in refining medical image segmentation, highlighting their potential to significantly boost model accuracy and robustness. The source code and our implementation are available at: https://bit.ly/3zf2CVs

comment: Submitted to IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025

☆ Algorithms For Automatic Accentuation And Transcription Of Russian Texts In Speech Recognition Systems SP

Olga Iakovenko, Ivan Bondarenko, Mariya Borovikova, Daniil Vodolazsky

This paper presents an overview of rule-based system for automatic accentuation and phonemic transcription of Russian texts for speech connected tasks, such as Automatic Speech Recognition (ASR). Two parts of the developed system, accentuation and transcription, use different approaches to achieve correct phonemic representations of input phrases. Accentuation is based on "Grammatical dictionary of the Russian language" of A.A. Zaliznyak and wiktionary corpus. To distinguish homographs, the accentuation system also utilises morphological information of the sentences based on Recurrent Neural Networks (RNN). Transcription algorithms apply the rules presented in the monograph of B.M. Lobanov and L.I. Tsirulnik "Computer Synthesis and Voice Cloning". The rules described in the present paper are implemented in an open-source module, which can be of use to any scientific study connected to ASR or Speech To Text (STT) tasks. Automatically marked up text annotations of the Russian Voxforge database were used as training data for an acoustic model in CMU Sphinx. The resulting acoustic model was evaluated on cross-validation, mean Word Accuracy being 71.2%. The developed toolkit is written in the Python language and is accessible on GitHub for any researcher interested.

comment: Speech and Computer 20th International Conference, SPECOM 2018, Leipzig, Germany, Proceedings 20

☆ Contextual Document Embeddings

John X. Morris, Alexander M. Rush

☆ Methods for Automatic Matrix Language Determination of Code-Switched Speech EMNLP

Olga Iakovenko, Thomas Hain

Code-switching (CS) is the process of speakers interchanging between two or more languages which in the modern world becomes increasingly common. In order to better describe CS speech the Matrix Language Frame (MLF) theory introduces the concept of a Matrix Language, which is the language that provides the grammatical structure for a CS utterance. In this work the MLF theory was used to develop systems for Matrix Language Identity (MLID) determination. The MLID of English/Mandarin and English/Spanish CS text and speech was compared to acoustic language identity (LID), which is a typical way to identify a language in monolingual utterances. MLID predictors from audio show higher correlation with the textual principles than LID in all cases while also outperforming LID in an MLID recognition task based on F1 macro (60\%) and correlation score (0.38). This novel approach has identified that non-English languages (Mandarin and Spanish) are preferred over the English language as the ML contrary to the monolingual choice of LID.

comment: Accepted at EMNLP

☆ Can Large Language Models Grasp Legal Theories? Enhance Legal Reasoning with Insights from Multi-Agent Collaboration

Weikang Yuan, Junjie Cao, Zhuoren Jiang, Yangyang Kang, Jun Lin, Kaisong Song, tianqianjin lin, Pengwei Yan, Changlong Sun, Xiaozhong Liu

Large Language Models (LLMs) could struggle to fully understand legal theories and perform complex legal reasoning tasks. In this study, we introduce a challenging task (confusing charge prediction) to better evaluate LLMs' understanding of legal theories and reasoning capabilities. We also propose a novel framework: Multi-Agent framework for improving complex Legal Reasoning capability (MALR). MALR employs non-parametric learning, encouraging LLMs to automatically decompose complex legal tasks and mimic human learning process to extract insights from legal rules, helping LLMs better understand legal theories and enhance their legal reasoning abilities. Extensive experiments on multiple real-world datasets demonstrate that the proposed framework effectively addresses complex reasoning issues in practical scenarios, paving the way for more reliable applications in the legal domain.

☆ Mixed-Session Conversation with Egocentric Memory EMNLP

Jihyoung Jang, Taeyoung Kim, Hyounghun Kim

Recently introduced dialogue systems have demonstrated high usability. However, they still fall short of reflecting real-world conversation scenarios. Current dialogue systems exhibit an inability to replicate the dynamic, continuous, long-term interactions involving multiple partners. This shortfall arises because there have been limited efforts to account for both aspects of real-world dialogues: deeply layered interactions over the long-term dialogue and widely expanded conversation networks involving multiple participants. As the effort to incorporate these aspects combined, we introduce Mixed-Session Conversation, a dialogue system designed to construct conversations with various partners in a multi-session dialogue setup. We propose a new dataset called MiSC to implement this system. The dialogue episodes of MiSC consist of 6 consecutive sessions, with four speakers (one main speaker and three partners) appearing in each episode. Also, we propose a new dialogue model with a novel memory management mechanism, called Egocentric Memory Enhanced Mixed-Session Conversation Agent (EMMA). EMMA collects and retains memories from the main speaker's perspective during conversations with partners, enabling seamless continuity in subsequent interactions. Extensive human evaluations validate that the dialogues in MiSC demonstrate a seamless conversational flow, even when conversation partners change in each session. EMMA trained with MiSC is also evaluated to maintain high memorability without contradiction throughout the entire conversation.

comment: EMNLP Findings 2024 (30 pages); Project website: https://mixed-session.github.io/

☆ Defining Knowledge: Bridging Epistemology and Large Language Models EMNLP 2024

Constanza Fierro, Ruchira Dhar, Filippos Stamatiou, Nicolas Garneau, Anders Søgaard

Knowledge claims are abundant in the literature on large language models (LLMs); but can we say that GPT-4 truly "knows" the Earth is round? To address this question, we review standard definitions of knowledge in epistemology and we formalize interpretations applicable to LLMs. In doing so, we identify inconsistencies and gaps in how current NLP research conceptualizes knowledge with respect to epistemological frameworks. Additionally, we conduct a survey of 100 professional philosophers and computer scientists to compare their preferences in knowledge definitions and their views on whether LLMs can really be said to know. Finally, we suggest evaluation protocols for testing knowledge in accordance to the most relevant definitions.

comment: EMNLP 2024

☆ Dynamic Gradient Alignment for Online Data Mixing

Simin Fan, David Grangier, Pierre Ablin

The composition of training data mixtures is critical for effectively training large language models (LLMs), as it directly impacts their performance on downstream tasks. Our goal is to identify an optimal data mixture to specialize an LLM for a specific task with access to only a few examples. Traditional approaches to this problem include ad-hoc reweighting methods, importance sampling, and gradient alignment techniques. This paper focuses on gradient alignment and introduces Dynamic Gradient Alignment (DGA), a scalable online gradient alignment algorithm. DGA dynamically estimates the pre-training data mixture on which the models' gradients align as well as possible with those of the model on the specific task. DGA is the first gradient alignment approach that incurs minimal overhead compared to standard pre-training and outputs a competitive model, eliminating the need for retraining the model. Experimentally, we demonstrate significant improvements over importance sampling in two key scenarios: (i) when the pre-training set is small and importance sampling overfits due to limited data; and (ii) when there is insufficient specialized data, trapping importance sampling on narrow pockets of data. Our findings underscore the effectiveness of gradient alignment methods in optimizing training data mixtures, particularly in data-constrained environments, and offer a practical solution for enhancing LLM performance on specific tasks with limited data availability.

☆ DTVLT: A Multi-modal Diverse Text Benchmark for Visual Language Tracking Based on LLM

Xuchen Li, Shiyu Hu, Xiaokun Feng, Dailing Zhang, Meiqi Wu, Jing Zhang, Kaiqi Huang

Visual language tracking (VLT) has emerged as a cutting-edge research area, harnessing linguistic data to enhance algorithms with multi-modal inputs and broadening the scope of traditional single object tracking (SOT) to encompass video understanding applications. Despite this, most VLT benchmarks still depend on succinct, human-annotated text descriptions for each video. These descriptions often fall short in capturing the nuances of video content dynamics and lack stylistic variety in language, constrained by their uniform level of detail and a fixed annotation frequency. As a result, algorithms tend to default to a "memorize the answer" strategy, diverging from the core objective of achieving a deeper understanding of video content. Fortunately, the emergence of large language models (LLMs) has enabled the generation of diverse text. This work utilizes LLMs to generate varied semantic annotations (in terms of text lengths and granularities) for representative SOT benchmarks, thereby establishing a novel multi-modal benchmark. Specifically, we (1) propose a new visual language tracking benchmark with diverse texts, named DTVLT, based on five prominent VLT and SOT benchmarks, including three sub-tasks: short-term tracking, long-term tracking, and global instance tracking. (2) We offer four granularity texts in our benchmark, considering the extent and density of semantic information. We expect this multi-granular generation strategy to foster a favorable environment for VLT and video understanding research. (3) We conduct comprehensive experimental analyses on DTVLT, evaluating the impact of diverse text on tracking performance and hope the identified performance bottlenecks of existing algorithms can support further research in VLT and video understanding. The proposed benchmark, experimental results and toolkit will be released gradually on http://videocube.aitestunion.com/.

comment: Preprint, Under Review

☆ Response Tuning: Aligning Large Language Models without Instruction

Seokhyun An, Hyounghun Kim

Instruction tuning-supervised fine-tuning using instruction-response pairs-is a foundational step in transitioning pre-trained Large Language Models (LLMs) into helpful and safe chat assistants. Our hypothesis is that establishing an adequate output space can enable such a transition given the capabilities inherent in pre-trained LLMs. To verify this, we propose Response Tuning (RT), which eliminates the instruction-conditioning step in instruction tuning and solely focuses on response space supervision. Our experiments demonstrate that RT models, trained only using responses, can effectively respond to a wide range of instructions and exhibit helpfulness comparable to that of their instruction-tuned counterparts. Furthermore, we observe that controlling the training response distribution can significantly improve their user preference or elicit target behaviors such as refusing assistance for unsafe queries. Our findings illuminate the role of establishing an adequate output space in alignment, highlighting the potential of the extensive inherent capabilities of pre-trained LLMs.

comment: 34 pages

☆ Embedded Topic Models Enhanced by Wikification EMNLP 2024

Takashi Shibuya, Takehito Utsuro

Topic modeling analyzes a collection of documents to learn meaningful patterns of words. However, previous topic models consider only the spelling of words and do not take into consideration the homography of words. In this study, we incorporate the Wikipedia knowledge into a neural topic model to make it aware of named entities. We evaluate our method on two datasets, 1) news articles of \textit{New York Times} and 2) the AIDA-CoNLL dataset. Our experiments show that our method improves the performance of neural topic models in generalizability. Moreover, we analyze frequent terms in each topic and the temporal dependencies between topics to demonstrate that our entity-aware topic models can capture the time-series development of topics well.

comment: Accepted at EMNLP 2024 Workshop NLP for Wikipedia

☆ Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization

Mingyang Wang, Lukas Lange, Heike Adel, Jannik Strötgen, Hinrich Schütze

To ensure large language models contain up-to-date knowledge, they need to be updated regularly. However, model editing is challenging as it might also affect knowledge that is unrelated to the new data. State-of-the-art methods identify parameters associated with specific knowledge and then modify them via direct weight updates. However, these locate-and-edit methods suffer from heavy computational overhead and lack theoretical validation. In contrast, directly fine-tuning the model on requested edits affects the model's behavior on unrelated knowledge, and significantly damages the model's generation fluency and consistency. To address these challenges, we propose SAUL, a streamlined model editing method that uses sentence concatenation with augmented random facts for generation regularization. Evaluations on three model editing benchmarks show that SAUL is a practical and reliable solution for model editing outperforming state-of-the-art methods while maintaining generation quality and reducing computational overhead.

☆ IoT-LLM: Enhancing Real-World IoT Task Reasoning with Large Language Models ICLR 2025

Tuo An, Yunjiao Zhou, Han Zou, Jianfei Yang

comment: 21 pages, 10 figures, submitted to ICLR 2025 Conference

☆ Collective Critics for Creative Story Generation EMNLP 2024

Minwook Bae, Hyounghun Kim

Generating a long story of several thousand words with narrative coherence using Large Language Models (LLMs) has been a challenging task. Previous research has addressed this challenge by proposing different frameworks that create a story plan and generate a long story based on that plan. However, these frameworks have been mainly focusing on maintaining narrative coherence in stories, often overlooking creativity in story planning and the expressiveness of the stories generated from those plans, which are desirable properties to captivate readers' interest. In this paper, we propose Collective Critics for Creative Story Generation framework (CritiCS), which is composed of plan refining stage (CrPlan) and story generation stage (CrText), to integrate a collective revision mechanism that promotes those properties into long-form story generation process. Specifically, in each stage, a group of LLM critics and one leader collaborate to incrementally refine drafts of plan and story throughout multiple rounds. Extensive human evaluation shows that the CritiCS can significantly enhance story creativity and reader engagement, while also maintaining narrative coherence. Furthermore, the design of the framework allows active participation from human writers in any role within the critique process, enabling interactive human-machine collaboration in story writing.

comment: EMNLP 2024 (36 pages)

☆ Learning the Latent Rules of a Game from Data: A Chess Story

Ben Fauber

We demonstrate that small pretrained foundational generative language models with millions of parameters can learn the latent rules of a process from data associated with the process. Inspired by Stefan Zweig's novella "Schachnovelle," also known as "The Royal Game" in English, we show that 28M and 125M parameter pretrained foundational small language models (SLMs) can be instruction fine-tuned with 1,000-to-1,000,000 examples to learn the rules of chess, propose legal moves, and accurately solve chess problems. We also explore the impact of successive language model fine-tuning epochs on improved outcomes and demonstrate reductions in model hallucinations by increasing the number of instruction fine-tuning examples.

☆ LLM-Pilot: Characterize and Optimize Performance of your LLM Inference Services SC '24

Małgorzata Łazuka, Andreea Anghel, Thomas Parnell

comment: Accepted to the International Conference for High Performance Computing, Networking, Storage and Analysis (SC '24)

☆ MenakBERT -- Hebrew Diacriticizer SC

Ido Cohen, Jacob Gidron, Idan Pinto

Diacritical marks in the Hebrew language give words their vocalized form. The task of adding diacritical marks to plain Hebrew text is still dominated by a system that relies heavily on human-curated resources. Recent models trained on diacritized Hebrew texts still present a gap in performance. We use a recently developed char-based PLM to narrowly bridge this gap. Presenting MenakBERT, a character level transformer pretrained on Hebrew text and fine-tuned to produce diacritical marks for Hebrew sentences. We continue to show how finetuning a model for diacritizing transfers to a task such as part of speech tagging.

comment: Published at ISCOL2022 as a poster

☆ Parameter Competition Balancing for Model Merging NeurIPS2024

Guodong Du, Junlin Lee, Jing Li, Runhua Jiang, Yifei Guo, Shuyang Yu, Hanting Liu, Sim Kuan Goh, Ho-Kin Tang, Daojing He, Min Zhang

While fine-tuning pretrained models has become common practice, these models often underperform outside their specific domains. Recently developed model merging techniques enable the direct integration of multiple models, each fine-tuned for distinct tasks, into a single model. This strategy promotes multitasking capabilities without requiring retraining on the original datasets. However, existing methods fall short in addressing potential conflicts and complex correlations between tasks, especially in parameter-level adjustments, posing a challenge in effectively balancing parameter competition across various tasks. This paper introduces an innovative technique named PCB-Merging (Parameter Competition Balancing), a lightweight and training-free technique that adjusts the coefficients of each parameter for effective model merging. PCB-Merging employs intra-balancing to gauge parameter significance within individual tasks and inter-balancing to assess parameter similarities across different tasks. Parameters with low importance scores are dropped, and the remaining ones are rescaled to form the final merged model. We assessed our approach in diverse merging scenarios, including cross-task, cross-domain, and cross-training configurations, as well as out-of-domain generalization. The experimental results reveal that our approach achieves substantial performance enhancements across multiple modalities, domains, model sizes, number of tasks, fine-tuning forms, and large language models, outperforming existing model merging methods. The code is publicly available at: \url{https://github.com/duguodong7/pcb-merging}.

comment: Accepted by NeurIPS2024

☆ MetaMetrics: Calibrating Metrics For Generation Tasks Using Human Preferences

Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, Derry Tanti Wijaya

comment: Preprint

☆ Towards Comprehensive Detection of Chinese Harmful Memes

Junyu Lu, Bo Xu, Xiaokun Zhang, Hongbo Wang, Haohao Zhu, Dongyu Zhang, Liang Yang, Hongfei Lin

This paper has been accepted in the NeurIPS 2024 D & B Track. Harmful memes have proliferated on the Chinese Internet, while research on detecting Chinese harmful memes significantly lags behind due to the absence of reliable datasets and effective detectors. To this end, we focus on the comprehensive detection of Chinese harmful memes. We construct ToxiCN MM, the first Chinese harmful meme dataset, which consists of 12,000 samples with fine-grained annotations for various meme types. Additionally, we propose a baseline detector, Multimodal Knowledge Enhancement (MKE), incorporating contextual information of meme content generated by the LLM to enhance the understanding of Chinese memes. During the evaluation phase, we conduct extensive quantitative experiments and qualitative analyses on multiple baselines, including LLMs and our MKE. The experimental results indicate that detecting Chinese harmful memes is challenging for existing models while demonstrating the effectiveness of MKE. The resources for this paper are available at https://github.com/DUT-lujunyu/ToxiCN_MM.

☆ From Concrete to Abstract: A Multimodal Generative Approach to Abstract Concept Learning

Haodong Xie, Rahul Singh Maharjan, Federico Tavella, Angelo Cangelosi

Understanding and manipulating concrete and abstract concepts is fundamental to human intelligence. Yet, they remain challenging for artificial agents. This paper introduces a multimodal generative approach to high order abstract concept learning, which integrates visual and categorical linguistic information from concrete ones. Our model initially grounds subordinate level concrete concepts, combines them to form basic level concepts, and finally abstracts to superordinate level concepts via the grounding of basic-level concepts. We evaluate the model language learning ability through language-to-visual and visual-to-language tests with high order abstract concepts. Experimental results demonstrate the proficiency of the model in both language understanding and language naming tasks.

☆ AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models

Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Xiang Wang, Xiangnan He, Tat-seng Chua

Large language models (LLMs) often exhibit hallucinations due to incorrect or outdated knowledge. Hence, model editing methods have emerged to enable targeted knowledge updates. To achieve this, a prevailing paradigm is the locating-then-editing approach, which first locates influential parameters and then edits them by introducing a perturbation. While effective, current studies have demonstrated that this perturbation inevitably disrupt the originally preserved knowledge within LLMs, especially in sequential editing scenarios. To address this, we introduce AlphaEdit, a novel solution that projects perturbation onto the null space of the preserved knowledge before applying it to the parameters. We theoretically prove that this projection ensures the output of post-edited LLMs remains unchanged when queried about the preserved knowledge, thereby mitigating the issue of disruption. Extensive experiments on various LLMs, including LLaMA3, GPT2-XL, and GPT-J, show that AlphaEdit boosts the performance of most locating-then-editing methods by an average of 36.4% with a single line of additional code for projection solely. Our code is available at: https://github.com/jianghoucheng/AlphaEdit.

☆ Listening to the Wise Few: Select-and-Copy Attention Heads for Multiple-Choice QA

Eduard Tulchinskii, Laida Kushnareva, Kristian Kuznetsov, Anastasia Voznyuk, Andrei Andriiainen, Irina Piontkovskaya, Evgeny Burnaev, Serguei Barannikov

A standard way to evaluate the abilities of LLM involves presenting a multiple-choice question and selecting the option with the highest logit as the model's predicted answer. However, such a format for evaluating LLMs has limitations, since even if the model knows the correct answer, it may struggle to select the corresponding letter simply due to difficulties in following this rigid format. To address this, we introduce new scores that better capture and reveal model's underlying knowledge: the Query-Key Score (QK-score), derived from the interaction between query and key representations in attention heads, and the Attention Score, based on attention weights. These scores are extracted from specific \textit{select-and-copy} heads, which show consistent performance across popular Multi-Choice Question Answering (MCQA) datasets. Based on these scores, our method improves knowledge extraction, yielding up to 16\% gain for LLaMA2-7B and up to 10\% for larger models on popular MCQA benchmarks. At the same time, the accuracy on a simple synthetic dataset, where the model explicitly knows the right answer, increases by almost 60\%, achieving nearly perfect accuracy, therefore demonstrating the method's efficiency in mitigating MCQA format limitations. To support our claims, we conduct experiments on models ranging from 7 billion to 70 billion parameters in both zero- and few-shot setups.

☆ How Much Can RAG Help the Reasoning of LLM?

Jingyu Liu, Jiaen Lin, Yong Liu

☆ Llama SLayer 8B: Shallow Layers Hold the Key to Knowledge Injection

Tianxiang Chen, Zhentao Tan, Tao Gong, Yue Wu, Qi Chu, Bin Liu, Jieping Ye, Nenghai Yu

As a manner to augment pre-trained large language models (LLM), knowledge injection is critical to develop vertical domain large models and has been widely studied. Although most current approaches, including parameter-efficient fine-tuning (PEFT) and block expansion methods, uniformly apply knowledge across all LLM layers, it raises the question: are all layers equally crucial for knowledge injection? We begin by evaluating the importance of each layer in finding the optimal layer range for knowledge injection. Intuitively, the more important layers should play a more critical role in knowledge injection and deserve a denser injection. We observe performance dips in question-answering benchmarks after the removal or expansion of the shallow layers, and the degradation shrinks as the layer gets deeper, indicating that the shallow layers hold the key to knowledge injection. This insight leads us to propose the S strategy, a post-pretraining strategy of selectively enhancing shallow layers while pruning the less effective deep ones. Based on this strategy, we introduce Llama Slayer-8B and Llama Slayer-8B-Instruct. We experimented on the corpus of code $\&$ math and demonstrated the effectiveness of our strategy. Further experiments across different LLM, Mistral-7B, and a legal corpus confirmed the general applicability of the approach, underscoring its wide-ranging efficacy. Our code is available at: \https://github.com/txchen-USTC/Llama-Slayer

☆ Post-edits Are Preferences Too

Nathaniel Berger, Stefan Riezler, Miriam Exel, Matthias Huck

Preference Optimization (PO) techniques are currently one of the state of the art techniques for fine-tuning large language models (LLMs) on pairwise preference feedback from human annotators. However, in machine translation, this sort of feedback can be difficult to solicit. Additionally, Kreutzer et al. (2018) have shown that, for machine translation, pairwise preferences are less reliable than other forms of human feedback, such as 5-point ratings. We examine post-edits to see if they can be a source of reliable human preferences by construction. In PO, a human annotator is shown sequences $s_1$ and $s_2$ and asked for a preference judgment, %$s_1 > s_2$; while for post-editing, editors \emph{create} $s_1$ and know that it should be better than $s_2$. We attempt to use these implicit preferences for PO and show that it helps the model move towards post-edit-like hypotheses and away from machine translation-like hypotheses. Furthermore, we show that best results are obtained by pre-training the model with supervised fine-tuning (SFT) on post-edits in order to promote post-edit-like hypotheses to the top output ranks.

comment: To appear at the Ninth Conference on Machine Translation (WMT24)

☆ Traffic Light or Light Traffic? Investigating Phrasal Semantics in Large Language Models EMNLP 2024

Rui Meng, Ye Liu, Lifu Tu, Daqing He, Yingbo Zhou, Semih Yavuz

Phrases are fundamental linguistic units through which humans convey semantics. This study critically examines the capacity of API-based large language models (LLMs) to comprehend phrase semantics, utilizing three human-annotated datasets. We assess the performance of LLMs in executing phrase semantic reasoning tasks guided by natural language instructions and explore the impact of common prompting techniques, including few-shot demonstrations and Chain-of-Thought reasoning. Our findings reveal that LLMs greatly outperform traditional embedding methods across the datasets; however, they do not show a significant advantage over fine-tuned methods. The effectiveness of advanced prompting strategies shows variability. We conduct detailed error analyses to interpret the limitations faced by LLMs in comprehending phrase semantics. Code and data can be found at https://github.com/memray/llm_phrase_semantics.

comment: EMNLP 2024

☆ Jailbreak Antidote: Runtime Safety-Utility Balance via Sparse Representation Adjustment in Large Language Models

Guobin Shen, Dongcheng Zhao, Yiting Dong, Xiang He, Yi Zeng

comment: 10 pages, 5 figures

☆ Make Compound Sentences Simple to Analyze: Learning to Split Sentences for Aspect-based Sentiment Analysis EMNLP 2024

Yongsik Seo, Sungwon Song, Ryang Heo, Jieyong Kim, Dongha Lee

In the domain of Aspect-Based Sentiment Analysis (ABSA), generative methods have shown promising results and achieved substantial advancements. However, despite these advancements, the tasks of extracting sentiment quadruplets, which capture the nuanced sentiment expressions within a sentence, remain significant challenges. In particular, compound sentences can potentially contain multiple quadruplets, making the extraction task increasingly difficult as sentence complexity grows. To address this issue, we are focusing on simplifying sentence structures to facilitate the easier recognition of these elements and crafting a model that integrates seamlessly with various ABSA tasks. In this paper, we propose Aspect Term Oriented Sentence Splitter (ATOSS), which simplifies compound sentence into simpler and clearer forms, thereby clarifying their structure and intent. As a plug-and-play module, this approach retains the parameters of the ABSA model while making it easier to identify essential intent within input sentences. Extensive experimental results show that utilizing ATOSS outperforms existing methods in both ASQP and ACOS tasks, which are the primary tasks for extracting sentiment quadruplets.

comment: Accepted at EMNLP 2024 (Findings, long paper)

☆ Language Models are Graph Learners

Zhe Xu, Kaveh Hassani, Si Zhang, Hanqing Zeng, Michihiro Yasunaga, Limei Wang, Dongqi Fu, Ning Yao, Bo Long, Hanghang Tong

Language Models (LMs) are increasingly challenging the dominance of domain-specific models, including Graph Neural Networks (GNNs) and Graph Transformers (GTs), in graph learning tasks. Following this trend, we propose a novel approach that empowers off-the-shelf LMs to achieve performance comparable to state-of-the-art GNNs on node classification tasks, without requiring any architectural modification. By preserving the LM's original architecture, our approach retains a key benefit of LM instruction tuning: the ability to jointly train on diverse datasets, fostering greater flexibility and efficiency. To achieve this, we introduce two key augmentation strategies: (1) Enriching LMs' input using topological and semantic retrieval methods, which provide richer contextual information, and (2) guiding the LMs' classification process through a lightweight GNN classifier that effectively prunes class candidates. Our experiments on real-world datasets show that backbone Flan-T5 models equipped with these augmentation strategies outperform state-of-the-art text-output node classifiers and are comparable to top-performing vector-output node classifiers. By bridging the gap between specialized task-specific node classifiers and general LMs, this work paves the way for more versatile and widely applicable graph learning models. We will open-source the code upon publication.

☆ Efficient Second-Order Neural Network Optimization via Adaptive Trust Region Methods

James Vo

Second-order optimization methods offer notable advantages in training deep neural networks by utilizing curvature information to achieve faster convergence. However, traditional second-order techniques are computationally prohibitive, primarily due to the large matrix inversions and high memory demands they require. While adaptive trust-region methods have been developed to mitigate these issues, their performance is often hindered by conservative estimates of key parameters, such as the Lipschitz constant of the Hessian, resulting in suboptimal outcomes. In this paper, we introduce SecondOrderAdaptiveAdam (SOAA), a novel optimization algorithm designed to overcome these limitations. SOAA approximates the Fisher information matrix using a diagonal representation, reducing computational complexity from $O(n^{2})$ to $O(n)$, thereby making it suitable for large-scale deep learning models, including large language models (LLMs). Additionally, the algorithm integrates an adaptive trust-region mechanism that dynamically adjusts the trust region size based on observed loss reduction, ensuring both robust convergence and computational efficiency. We empirically demonstrate that SOAA achieves faster and more stable convergence compared to first-order optimizers, such as Adam, under similar computational constraints. However, the diagonal approximation of the Fisher information matrix may be less effective in capturing higher-order interactions between gradients, suggesting potential areas for further refinement and future research.

☆ Correlation and Navigation in the Vocabulary Key Representation Space of Language Models

Letian Peng, Chenyang An, Jingbo Shang

Language model (LM) decoding is based on the next-token prediction (NTP) probability distribution. For neural LMs (e.g., Transformer-based), NTP distribution is essentially a softmax-regularized dot product between an encoded input context (query) and fixed vocabulary representations (keys). In this paper, we study the effect of the key distribution on the NTP distribution, with a focus on whether the similarity between keys will trigger spurious correlations in NTP. Through knowledge-probing tasks, we show that in the NTP distribution, the few top-ranked tokens are typically accurate. However, the middle-ranked prediction is highly biased towards the tokens that are distributionally (not necessarily semantically) similar to these top ones. For instance, if "P" is predicted as the top-1 token, "A"-"Z" will all be ranked high in NTP, no matter whether they can lead to correct decoding results. This hurts the sampling diversity and makes the sampling of correct, long-tail results hopeless and noisy. We attempt to alleviate this issue via a novel in-context method that iteratively pushes the query representation away from explored regions. Specifically, we include the explored decoding results in the context and prompt the LM to generate something else, which encourages the LM to produce a query representation that has small dot products with explored keys. Experiments on knowledge-probing tasks show that our method leads to efficient navigation away from explored keys to correct new keys. We further extend our method to open-ended and chain-of-thought (for reasoning) generation. Experiment results show that ICN contributes to better generation diversity and improved self-consistency voting performance. Finally, we discuss potential training issues caused by the fixed key space together with the challenges and possible ways to address them in future research.

☆ Morphological evaluation of subwords vocabulary used by BETO language model

Óscar García-Sierra, Ana Fernández-Pampillón Cesteros, Miguel Ortega-Martín

Subword tokenization algorithms used by Large Language Models are significantly more efficient and can independently build the necessary vocabulary of words and subwords without human intervention. However, those subwords do not always align with real morphemes, potentially impacting the models' performance, though it remains uncertain when this might occur. In previous research, we proposed a method to assess the morphological quality of vocabularies, focusing on the overlap between these vocabularies and the morphemes of a given language. Our evaluation method was built on three quality measures, relevance, cohesion, and morphological accuracy, and a procedure for their assessment. By applying this method to vocabularies created by three subword tokenization algorithms, BPE, Wordpiece, and Unigram, we concluded that these vocabularies generally exhibit very low morphological quality. In this article, we apply this evaluation to the tokenizer of BETO, a BERT language model trained on large Spanish corpora. This evaluation, along with our previous results, helped us conclude that its vocabulary has a low morphological quality, and we also found that training the tokenizer in a larger corpus does not improve the morphological quality of the generated vocabulary. Additionally, this evaluation helps clarify the algorithm used by the tokenizer, that is, Wordpiece, given the inconsistencies between the authors' claims and the model's configuration.

comment: in Spanish language

☆ Annotation Guidelines for Corpus Novelties: Part 1 -- Named Entity Recognition

Arthur Amalvy, Vincent Labatut

♻ ☆ Which questions should I answer? Salience Prediction of Inquisitive Questions EMNLP 2024

Yating Wu, Ritika Mangla, Alexandros G. Dimakis, Greg Durrett, Junyi Jessy Li

Inquisitive questions -- open-ended, curiosity-driven questions people ask as they read -- are an integral part of discourse processing (Kehler and Rohde, 2017; Onea, 2016) and comprehension (Prince, 2004). Recent work in NLP has taken advantage of question generation capabilities of LLMs to enhance a wide range of applications. But the space of inquisitive questions is vast: many questions can be evoked from a given context. So which of those should be prioritized to find answers? Linguistic theories, unfortunately, have not yet provided an answer to this question. This paper presents QSALIENCE, a salience predictor of inquisitive questions. QSALIENCE is instruction-tuned over our dataset of linguist-annotated salience scores of 1,766 (context, question) pairs. A question scores high on salience if answering it would greatly enhance the understanding of the text (Van Rooy, 2003). We show that highly salient questions are empirically more likely to be answered in the same article, bridging potential questions (Onea, 2016) with Questions Under Discussion (Roberts, 2012). We further validate our findings by showing that answering salient questions is an indicator of summarization quality in news.

comment: Camera Ready for EMNLP 2024 Main Conference

♻ ☆ LML-DAP: Language Model Learning a Dataset for Data-Augmented Prediction

Praneeth Vadlapati

Classification tasks are typically handled using Machine Learning (ML) models, which lack a balance between accuracy and interpretability. This paper introduces a new approach to using Large Language Models (LLMs) for classification tasks in an explainable way. Unlike ML models that rely heavily on data cleaning and feature engineering, this method streamlines the process using LLMs. This paper proposes a new concept called "Language Model Learning (LML)" powered by a new method called "Data-Augmented Prediction (DAP)". The classification is performed by LLMs using a method similar to humans manually exploring and understanding the data and deciding classifications using data as a reference. In the LML process, a dataset is summarized and evaluated to determine the features that lead to the classification of each label the most. In the process of DAP, the system uses the data summary and a row of the testing dataset to automatically generate a query, which is used to retrieve relevant rows from the dataset. A classification is generated by the LLM using data summary and relevant rows, ensuring satisfactory accuracy even with complex data using context-aware decision-making. LML and DAP unlock the possibilities of new applications. The proposed method uses the words "Act as an Explainable Machine Learning Model" in the prompt to enhance the interpretability of the predictions by allowing users to review the logic behind each prediction. In some test cases, the system scored an accuracy above 90%, proving the effectiveness of the system and its potential to outperform conventional ML models in various scenarios. The code is available at https://github.com/Pro-GenAI/LML-DAP

comment: Updated title, abstract, and images

♻ ☆ Tokenization Falling Short: The Curse of Tokenization EMNLP 2024

Yekun Chai, Yewei Fang, Qiwei Peng, Xuhong Li

comment: EMNLP 2024 Findings

♻ ☆ On Training Data Influence of GPT Models EMNLP 2024

Yekun Chai, Qingyi Liu, Shuohuan Wang, Yu Sun, Qiwei Peng, Hua Wu

Amidst the rapid advancements in generative language models, the investigation of how training data shapes the performance of GPT models is still emerging. This paper presents GPTfluence, a novel approach that leverages a featurized simulation to assess the impact of training examples on the training dynamics of GPT models. Our approach not only traces the influence of individual training instances on performance trajectories, such as loss and other key metrics, on targeted test points but also enables a comprehensive comparison with existing methods across various training scenarios in GPT models, ranging from 14 million to 2.8 billion parameters, across a range of downstream tasks. Contrary to earlier methods that struggle with generalization to new data, GPTfluence introduces a parameterized simulation of training dynamics, demonstrating robust generalization capabilities to unseen training data. This adaptability is evident across both fine-tuning and instruction-tuning scenarios, spanning tasks in natural language understanding and generation. We make our code and data publicly available at https://github.com/ernie-research/gptfluence.

comment: EMNLP 2024

♻ ☆ PharmacyGPT: The AI Pharmacist

Zhengliang Liu, Zihao Wu, Mengxuan Hu, Bokai Zhao, Lin Zhao, Tianyi Zhang, Haixing Dai, Xianyan Chen, Ye Shen, Sheng Li, Quanzheng Li, Xiang Li, Brian Murray, Tianming Liu, Andrea Sikora

In this study, we introduce PharmacyGPT, a novel framework to assess the capabilities of large language models (LLMs) such as ChatGPT and GPT-4 in emulating the role of clinical pharmacists. Our methodology encompasses the utilization of LLMs to generate comprehensible patient clusters, formulate medication plans, and forecast patient outcomes. We conduct our investigation using real data acquired from the intensive care unit (ICU) at the University of North Carolina Chapel Hill (UNC) Hospital. Our analysis offers valuable insights into the potential applications and limitations of LLMs in the field of clinical pharmacy, with implications for both patient care and the development of future AI-driven healthcare solutions. By evaluating the performance of PharmacyGPT, we aim to contribute to the ongoing discourse surrounding the integration of artificial intelligence in healthcare settings, ultimately promoting the responsible and efficacious use of such technologies.

♻ ☆ Autoregressive Pre-Training on Pixels and Texts EMNLP 2024

Yekun Chai, Qingyi Liu, Jingwu Xiao, Shuohuan Wang, Yu Sun, Hua Wu

The integration of visual and textual information represents a promising direction in the advancement of language models. In this paper, we explore the dual modality of language--both visual and textual--within an autoregressive framework, pre-trained on both document images and texts. Our method employs a multimodal training strategy, utilizing visual data through next patch prediction with a regression head and/or textual data through next token prediction with a classification head. We focus on understanding the interaction between these two modalities and their combined impact on model performance. Our extensive evaluation across a wide range of benchmarks shows that incorporating both visual and textual data significantly improves the performance of pixel-based language models. Remarkably, we find that a unidirectional pixel-based model trained solely on visual data can achieve comparable results to state-of-the-art bidirectional models on several language understanding tasks. This work uncovers the untapped potential of integrating visual and textual modalities for more effective language modeling. We release our code, data, and model checkpoints at \url{https://github.com/ernie-research/pixelgpt}.

comment: EMNLP 2024

♻ ☆ Turning English-centric LLMs Into Polyglots: How Much Multilinguality Is Needed? EMNLP 2024

Tannon Kew, Florian Schottmann, Rico Sennrich

The vast majority of today's large language models (LLMs) are English-centric, having been pretrained predominantly on English text. Yet, in order to meet user expectations, models need to be able to respond appropriately in multiple languages once deployed in downstream applications. This requires strong cross-lingual transfer abilities. In this work, we investigate the minimal amount of multilinguality required during finetuning to elicit cross-lingual generalisation in English-centric LLMs. In experiments across four LLMs, we find that multilingual instruction tuning with as few as two to three languages is both necessary and sufficient to elicit effective cross-lingual generalisation, with the limiting factor being the degree to which a target language is seen during pretraining. Evaluations on five different tasks further reveal that multilingual instruction tuning is most beneficial for generative tasks that assume input/output language agreement, such as in chat settings, while being of less importance for highly structured classification-style tasks. Our code and data is available at https://github.com/ZurichNLP/multilingual-instruction-tuning.

comment: Accepted at Findings of EMNLP 2024

♻ ☆ Lookback Lens: Detecting and Mitigating Contextual Hallucinations in Large Language Models Using Only Attention Maps EMNLP 2024

Yung-Sung Chuang, Linlu Qiu, Cheng-Yu Hsieh, Ranjay Krishna, Yoon Kim, James Glass

When asked to summarize articles or answer questions given a passage, large language models (LLMs) can hallucinate details and respond with unsubstantiated answers that are inaccurate with respect to the input context. This paper describes a simple approach for detecting such contextual hallucinations. We hypothesize that contextual hallucinations are related to the extent to which an LLM attends to information in the provided context versus its own generations. Based on this intuition, we propose a simple hallucination detection model whose input features are given by the ratio of attention weights on the context versus newly generated tokens (for each attention head). We find that a linear classifier based on these lookback ratio features is as effective as a richer detector that utilizes the entire hidden states of an LLM or a text-based entailment model. The lookback ratio-based detector -- Lookback Lens -- is found to transfer across tasks and even models, allowing a detector that is trained on a 7B model to be applied (without retraining) to a larger 13B model. We further apply this detector to mitigate contextual hallucinations, and find that a simple classifier-guided decoding approach is able to reduce the amount of hallucination, for example by 9.6% in the XSum summarization task.

comment: EMNLP 2024 main conference long paper. The source code is available at https://github.com/voidism/Lookback-Lens

♻ ☆ The Mystery of In-Context Learning: A Comprehensive Survey on Interpretation and Analysis EMNLP 2024

Yuxiang Zhou, Jiazheng Li, Yanzheng Xiang, Hanqi Yan, Lin Gui, Yulan He

Understanding in-context learning (ICL) capability that enables large language models (LLMs) to excel in proficiency through demonstration examples is of utmost importance. This importance stems not only from the better utilization of this capability across various tasks, but also from the proactive identification and mitigation of potential risks, including concerns regarding truthfulness, bias, and toxicity, that may arise alongside the capability. In this paper, we present a thorough survey on the interpretation and analysis of in-context learning. First, we provide a concise introduction to the background and definition of in-context learning. Then, we give an overview of advancements from two perspectives: 1) a theoretical perspective, emphasizing studies on mechanistic interpretability and delving into the mathematical foundations behind ICL; and 2) an empirical perspective, concerning studies that empirically analyze factors associated with ICL. We conclude by highlighting the challenges encountered and suggesting potential avenues for future research. We believe that our work establishes the basis for further exploration into the interpretation of in-context learning. Additionally, we have created a repository containing the resources referenced in our survey.

comment: Accepted to the main conference of EMNLP 2024. Resources are available at https://github.com/zyxnlp/ICL-Interpretation-Analysis-Resources

♻ ☆ Enhanced Automated Code Vulnerability Repair using Large Language Models

David de-Fitero-Dominguez, Eva Garcia-Lopez, Antonio Garcia-Cabot, Jose-Javier Martinez-Herraiz

This research addresses the complex challenge of automated repair of code vulnerabilities, vital for enhancing digital security in an increasingly technology-driven world. The study introduces a novel and efficient format for the representation of code modification, using advanced Large Language Models (LLMs) such as Code Llama and Mistral. These models, fine-tuned on datasets featuring C code vulnerabilities, significantly improve the accuracy and adaptability of automated code repair techniques. A key finding is the enhanced repair accuracy of these models when compared to previous methods such as VulRepair, which underscores their practical utility and efficiency. The research also offers a critical assessment of current evaluation metrics, such as perfect predictions, and their limitations in reflecting the true capabilities of automated repair models in real-world scenarios. Following this, it underscores the importance of using test datasets devoid of train samples, emphasizing the need for dataset integrity to enhance the effectiveness of LLMs in code repair tasks. The significance of this work is its contribution to digital security, setting new standards for automated code vulnerability repair and paving the way for future advancements in the fields of cybersecurity and artificial intelligence. The study does not only highlight the potential of LLMs in enhancing code security but also fosters further exploration and research in these crucial areas.

♻ ☆ On the Limited Generalization Capability of the Implicit Reward Model Induced by Direct Preference Optimization EMNLP

Yong Lin, Skyler Seto, Maartje ter Hoeve, Katherine Metcalf, Barry-John Theobald, Xuan Wang, Yizhe Zhang, Chen Huang, Tong Zhang

Reinforcement Learning from Human Feedback (RLHF) is an effective approach for aligning language models to human preferences. Central to RLHF is learning a reward function for scoring human preferences. Two main approaches for learning a reward model are 1) training an EXplicit Reward Model (EXRM) as in RLHF, and 2) using an implicit reward learned from preference data through methods such as Direct Preference Optimization (DPO). Prior work has shown that the implicit reward model of DPO (denoted as DPORM) can approximate an EXRM in the limit. DPORM's effectiveness directly implies the optimality of the learned policy, and also has practical implication for LLM alignment methods including iterative DPO. However, it is unclear how well DPORM empirically matches the performance of EXRM. This work studies the accuracy at distinguishing preferred and rejected answers for both DPORM and EXRM. Our findings indicate that even though DPORM fits the training dataset comparably, it generalizes less effectively than EXRM, especially when the validation datasets contain distribution shifts. Across five out-of-distribution settings, DPORM has a mean drop in accuracy of 3% and a maximum drop of 7%. These findings highlight that DPORM has limited generalization ability and substantiates the integration of an explicit reward model in iterative DPO approaches.

comment: 12 pages, 8 tables, 3 figures; Paper Accepted at EMNLP Findings 2024

♻ ☆ Jailbreaking LLMs with Arabic Transliteration and Arabizi EMNLP 2024

Mansour Al Ghanim, Saleh Almohaimeed, Mengxin Zheng, Yan Solihin, Qian Lou

This study identifies the potential vulnerabilities of Large Language Models (LLMs) to 'jailbreak' attacks, specifically focusing on the Arabic language and its various forms. While most research has concentrated on English-based prompt manipulation, our investigation broadens the scope to investigate the Arabic language. We initially tested the AdvBench benchmark in Standardized Arabic, finding that even with prompt manipulation techniques like prefix injection, it was insufficient to provoke LLMs into generating unsafe content. However, when using Arabic transliteration and chatspeak (or arabizi), we found that unsafe content could be produced on platforms like OpenAI GPT-4 and Anthropic Claude 3 Sonnet. Our findings suggest that using Arabic and its various forms could expose information that might remain hidden, potentially increasing the risk of jailbreak attacks. We hypothesize that this exposure could be due to the model's learned connection to specific words, highlighting the need for more comprehensive safety training across all language forms.

comment: Accepted by EMNLP 2024

♻ ☆ StorySparkQA: Expert-Annotated QA Pairs with Real-World Knowledge for Children's Story-Based Learning EMNLP 2024

Jiaju Chen, Yuxuan Lu, Shao Zhang, Bingsheng Yao, Yuanzhe Dong, Ying Xu, Yunyao Li, Qianwen Wang, Dakuo Wang, Yuling Sun

comment: Accepted at EMNLP 2024 Main Conference

♻ ☆ Rel-A.I.: An Interaction-Centered Approach To Measuring Human-LM Reliance

Kaitlyn Zhou, Jena D. Hwang, Xiang Ren, Nouha Dziri, Dan Jurafsky, Maarten Sap

The ability to communicate uncertainty, risk, and limitation is crucial for the safety of large language models. However, current evaluations of these abilities rely on simple calibration, asking whether the language generated by the model matches appropriate probabilities. Instead, evaluation of this aspect of LLM communication should focus on the behaviors of their human interlocutors: how much do they rely on what the LLM says? Here we introduce an interaction-centered evaluation framework called Rel-A.I. (pronounced "rely"}) that measures whether humans rely on LLM generations. We use this framework to study how reliance is affected by contextual features of the interaction (e.g, the knowledge domain that is being discussed), or the use of greetings communicating warmth or competence (e.g., "I'm happy to help!"). We find that contextual characteristics significantly affect human reliance behavior. For example, people rely 10% more on LMs when responding to questions involving calculations and rely 30% more on LMs that are perceived as more competent. Our results show that calibration and language quality alone are insufficient in evaluating the risks of human-LM interactions, and illustrate the need to consider features of the interactional context.

comment: Preprint

♻ ☆ On Leakage of Code Generation Evaluation Datasets EMNLP 2024

Alexandre Matton, Tom Sherborne, Dennis Aumiller, Elena Tommasone, Milad Alizadeh, Jingyi He, Raymond Ma, Maxime Voisin, Ellen Gilsenan-McMahon, Matthias Gallé

In this paper, we consider contamination by code generation test sets, in particular in their use in modern large language models. We discuss three possible sources of such contamination and show findings supporting each of them: (i) direct data leakage, (ii) indirect data leakage through the use of synthetic data and (iii) overfitting to evaluation sets during model selection. To address this, we release Less Basic Python Problems (LBPP): an uncontaminated new benchmark of 161 prompts with their associated Python solutions. LBPP is released at https://huggingface.co/datasets/CohereForAI/lbpp .

comment: EMNLP 2024 Findings. 5 main pages, 9 in total

♻ ☆ Does Refusal Training in LLMs Generalize to the Past Tense?

Maksym Andriushchenko, Nicolas Flammarion

Refusal training is widely used to prevent LLMs from generating harmful, undesirable, or illegal outputs. We reveal a curious generalization gap in the current refusal training approaches: simply reformulating a harmful request in the past tense (e.g., "How to make a Molotov cocktail?" to "How did people make a Molotov cocktail?") is often sufficient to jailbreak many state-of-the-art LLMs. We systematically evaluate this method on Llama-3 8B, Claude-3.5 Sonnet, GPT-3.5 Turbo, Gemma-2 9B, Phi-3-Mini, GPT-4o mini, GPT-4o, o1-mini, o1-preview, and R2D2 models using GPT-3.5 Turbo as a reformulation model. For example, the success rate of this simple attack on GPT-4o increases from 1% using direct requests to 88% using 20 past tense reformulation attempts on harmful requests from JailbreakBench with GPT-4 as a jailbreak judge. Interestingly, we also find that reformulations in the future tense are less effective, suggesting that refusal guardrails tend to consider past historical questions more benign than hypothetical future questions. Moreover, our experiments on fine-tuning GPT-3.5 Turbo show that defending against past reformulations is feasible when past tense examples are explicitly included in the fine-tuning data. Overall, our findings highlight that the widely used alignment techniques -- such as SFT, RLHF, and adversarial training -- employed to align the studied models can be brittle and do not always generalize as intended. We provide code and jailbreak artifacts at https://github.com/tml-epfl/llm-past-tense.

comment: Update in v3: o1-mini and o1-preview results (on top of GPT-4o and Claude 3.5 Sonnet added in v2). We provide code and jailbreak artifacts at https://github.com/tml-epfl/llm-past-tense

♻ ☆ Immunization against harmful fine-tuning attacks EMNLP 2024

Domenic Rosati, Jan Wehner, Kai Williams, Łukasz Bartoszcze, Jan Batzner, Hassan Sajjad, Frank Rudzicz

Large Language Models (LLMs) are often trained with safety guards intended to prevent harmful text generation. However, such safety training can be removed by fine-tuning the LLM on harmful datasets. While this emerging threat (harmful fine-tuning attacks) has been characterized by previous work, there is little understanding of how we should proceed in constructing and validating defenses against these attacks especially in the case where defenders would not have control of the fine-tuning process. We introduce a formal framework based on the training budget of an attacker which we call "Immunization" conditions. Using a formal characterisation of the harmful fine-tuning problem, we provide a thorough description of what a successful defense must comprise of and establish a set of guidelines on how rigorous defense research that gives us confidence should proceed.

comment: Published in EMNLP 2024

♻ ☆ Foundations of Large Language Model Compression -- Part 1: Weight Quantization

Sean I. Young

In recent years, compression of large language models (LLMs) has emerged as an important problem to enable language model deployment on resource-constrained devices, reduce computational costs, and mitigate the environmental footprint of large-scale AI infrastructure. In this paper, we lay down the foundation for LLM quantization from a convex optimization perspective and propose a quantization technique that builds on this foundation for optimum quantization outcomes. Our quantization framework, CVXQ, scales to models containing hundreds of billions of weight parameters and provides users with the flexibility to compress models to any specified model size, post-training. A reference implementation of CVXQ can be obtained from github.com/seannz/cvxq.

comment: Preprint. 17 pages, 4 figures, 5 appendices

♻ ☆ EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage

Zeyi Liao, Lingbo Mo, Chejian Xu, Mintong Kang, Jiawei Zhang, Chaowei Xiao, Yuan Tian, Bo Li, Huan Sun

comment: 29 pages

♻ ☆ PRompt Optimization in Multi-Step Tasks (PROMST): Integrating Human Feedback and Heuristic-based Sampling EMNLP 2024

Yongchao Chen, Jacob Arkin, Yilun Hao, Yang Zhang, Nicholas Roy, Chuchu Fan

Prompt optimization aims to find the best prompt to a large language model (LLM) for a given task. LLMs have been successfully used to help find and improve prompt candidates for single-step tasks. However, realistic tasks for agents are multi-step and introduce new challenges: (1) Prompt content is likely to be more extensive and complex, making it more difficult for LLMs to analyze errors, (2) the impact of an individual step is difficult to evaluate, and (3) different people may have varied preferences about task execution. While humans struggle to optimize prompts, they are good at providing feedback about LLM outputs; we therefore introduce a new LLM-driven discrete prompt optimization framework PRompt Optimization in Multi-Step Tasks (PROMST) that incorporates human-designed feedback rules to automatically offer direct suggestions for improvement. We also use an extra learned heuristic model that predicts prompt performance to efficiently sample from prompt candidates. This approach significantly outperforms both human-engineered prompts and several other prompt optimization methods across 11 representative multi-step tasks (an average 10.6\%-29.3\% improvement to current best methods on five LLMs respectively). We believe our work can serve as a benchmark for automatic prompt optimization for LLM-driven multi-step tasks. Datasets and Codes are available at https://github.com/yongchao98/PROMST. Project Page is available at https://yongchao98.github.io/MIT-REALM-PROMST.

comment: 62 pages, 14 figures, Published in EMNLP 2024 Main

♻ ☆ PARAMANU-AYN: Pretrain from scratch or Continual Pretraining of LLMs for Legal Domain Adaptation?

Mitodru Niyogi, Arnab Bhattacharya

In this paper, we present Paramanu-Ayn, a collection of legal language models trained exclusively on Indian legal case documents. This 97-million-parameter Auto-Regressive (AR) decoder-only model was pretrained from scratch with a context size of 8192 on a single GPU for just 185 hours, achieving an efficient MFU of 41.35. We also developed a legal domain specialized BPE tokenizer. We evaluated our model using perplexity and zero-shot tasks: case judgment prediction with explanation and abstractive case summarization. Paramanu-Ayn outperformed Llama-2 7B and Gemini-Pro in case judgment prediction with explanation task on test accuracy by nearly 2 percentage points, despite being 72 times smaller. In zero-shot abstractive summarization, it surpassed decoder-only LLMs generating fixed-length summaries (5000 tokens) by over 10 percentage points in BLEU and METEOR metrics, and by nearly 4 percentage points in BERTScore. Further evaluations on zero-shot commonsense and mathematical benchmarks showed that Paramanu-Ayn excelled despite being trained exclusively on legal documents, outperforming Llama-1, Llama-2, and Falcon on AGIEVAL-AQuA-RAT and AGIEVAL-SAT-Math tasks. We also instruction-tuned our model on 10,763 diverse legal tasks, including legal clause generation, legal drafting, case summarization, etc. The Paramanu-Ayn-instruct model scored above 8 out of 10 in clarity, relevance, completeness, and legal reasoning metrics by GPT-3.5-Turbo. We found that our models, were able to learn drafting knowledge and generalize to draft legal contracts and legal clauses with limited instruction-tuning. Hence, we conclude that for a strong domain-specialized generative language model (such as legal), domain specialized pretraining from scratch is more cost effective, environmentally friendly, and remains competitive with larger models or even better than adapting LLMs for legal domain tasks.

♻ ☆ Leopard: A Vision Language Model For Text-Rich Multi-Image Tasks

Mengzhao Jia, Wenhao Yu, Kaixin Ma, Tianqing Fang, Zhihan Zhang, Siru Ouyang, Hongming Zhang, Meng Jiang, Dong Yu

Text-rich images, where text serves as the central visual element guiding the overall understanding, are prevalent in real-world applications, such as presentation slides, scanned documents, and webpage snapshots. Tasks involving multiple text-rich images are especially challenging, as they require not only understanding the content of individual images but reasoning about inter-relationships and logical flows across multiple visual inputs. Despite the importance of these scenarios, current multimodal large language models (MLLMs) struggle to handle such tasks due to two key challenges: (1) the scarcity of high-quality instruction tuning datasets for text-rich multi-image scenarios, and (2) the difficulty in balancing image resolution with visual feature sequence length. To address these challenges, we propose Leopard, a MLLM designed specifically for handling vision-language tasks involving multiple text-rich images. First, we curated about one million high-quality multimodal instruction-tuning data, tailored to text-rich, multi-image scenarios. Second, we developed an adaptive high-resolution multi-image encoding module to dynamically optimize the allocation of visual sequence length based on the original aspect ratios and resolutions of the input images. Experiments across a wide range of benchmarks demonstrate our model's superior capabilities in text-rich, multi-image evaluations and competitive performance in general domain evaluations.

comment: Our code is available at https://github.com/Jill0001/Leopard

♻ ☆ Seemingly Plausible Distractors in Multi-Hop Reasoning: Are Large Language Models Attentive Readers?

Neeladri Bhuiya, Viktor Schlegel, Stefan Winkler

State-of-the-art Large Language Models (LLMs) are accredited with an increasing number of different capabilities, ranging from reading comprehension, over advanced mathematical and reasoning skills to possessing scientific knowledge. In this paper we focus on their multi-hop reasoning capability: the ability to identify and integrate information from multiple textual sources. Given the concerns with the presence of simplifying cues in existing multi-hop reasoning benchmarks, which allow models to circumvent the reasoning requirement, we set out to investigate, whether LLMs are prone to exploiting such simplifying cues. We find evidence that they indeed circumvent the requirement to perform multi-hop reasoning, but they do so in more subtle ways than what was reported about their fine-tuned pre-trained language model (PLM) predecessors. Motivated by this finding, we propose a challenging multi-hop reasoning benchmark, by generating seemingly plausible multi-hop reasoning chains, which ultimately lead to incorrect answers. We evaluate multiple open and proprietary state-of-the-art LLMs, and find that their performance to perform multi-hop reasoning is affected, as indicated by up to 45% relative decrease in F1 score when presented with such seemingly plausible alternatives. We conduct a deeper analysis and find evidence that while LLMs tend to ignore misleading lexical cues, misleading reasoning paths indeed present a significant challenge.

comment: 15 pages, 3 figures

♻ ☆ Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution

Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Yang Fan, Kai Dang, Mengfei Du, Xuancheng Ren, Rui Men, Dayiheng Liu, Chang Zhou, Jingren Zhou, Junyang Lin

We present the Qwen2-VL Series, an advanced upgrade of the previous Qwen-VL models that redefines the conventional predetermined-resolution approach in visual processing. Qwen2-VL introduces the Naive Dynamic Resolution mechanism, which enables the model to dynamically process images of varying resolutions into different numbers of visual tokens. This approach allows the model to generate more efficient and accurate visual representations, closely aligning with human perceptual processes. The model also integrates Multimodal Rotary Position Embedding (M-RoPE), facilitating the effective fusion of positional information across text, images, and videos. We employ a unified paradigm for processing both images and videos, enhancing the model's visual perception capabilities. To explore the potential of large multimodal models, Qwen2-VL investigates the scaling laws for large vision-language models (LVLMs). By scaling both the model size-with versions at 2B, 8B, and 72B parameters-and the amount of training data, the Qwen2-VL Series achieves highly competitive performance. Notably, the Qwen2-VL-72B model achieves results comparable to leading models such as GPT-4o and Claude3.5-Sonnet across various multimodal benchmarks, outperforming other generalist models. Code is available at https://github.com/QwenLM/Qwen2-VL .

comment: Code is available at https://github.com/QwenLM/Qwen2-VL. arXiv admin note: text overlap with arXiv:2408.15262 by other authors

♻ ☆ Fast Matrix Multiplications for Lookup Table-Quantized LLMs EMNLP 2024

Han Guo, William Brandon, Radostin Cholakov, Jonathan Ragan-Kelley, Eric P. Xing, Yoon Kim

comment: EMNLP 2024 (Findings)

♻ ☆ Llamipa: An Incremental Discourse Parser EMNLP 2024

Kate Thompson, Akshay Chaturvedi, Julie Hunter, Nicholas Asher

This paper provides the first discourse parsing experiments with a large language model(LLM) finetuned on corpora annotated in the style of SDRT (Segmented Discourse Representation Theory Asher, 1993; Asher and Lascarides, 2003). The result is a discourse parser, Llamipa (Llama Incremental Parser), that leverages discourse context, leading to substantial performance gains over approaches that use encoder-only models to provide local, context-sensitive representations of discourse units. Furthermore, it can process discourse data incrementally, which is essential for the eventual use of discourse information in downstream tasks.

comment: EMNLP 2024 Findings

♻ ☆ Nebula: A discourse aware Minecraft Builder EMNLP 2024

Akshay Chaturvedi, Kate Thompson, Nicholas Asher

When engaging in collaborative tasks, humans efficiently exploit the semantic structure of a conversation to optimize verbal and nonverbal interactions. But in recent "language to code" or "language to action" models, this information is lacking. We show how incorporating the prior discourse and nonlinguistic context of a conversation situated in a nonlinguistic environment can improve the "language to action" component of such interactions. We finetune an LLM to predict actions based on prior context; our model, Nebula, doubles the net-action F1 score over the baseline on this task of Jayannavar et al.(2020). We also investigate our model's ability to construct shapes and understand location descriptions using a synthetic dataset

comment: EMNLP 2024 Findings

♻ ☆ LongForm: Effective Instruction Tuning with Reverse Instructions EMNLP 2024

Abdullatif Köksal, Timo Schick, Anna Korhonen, Hinrich Schütze

Instruction tuning enables language models to more effectively generalize and better follow user intent. However, obtaining instruction data is costly and challenging. Prior work employs methods such as expensive human annotation, crowd-sourced datasets with alignment issues, and generating noisy examples via LLMs. We introduce the LongForm-C dataset, which is created by reverse instructions. We generate instructions via LLMs for human-written corpus examples using reverse instructions. First we select a diverse set of human-written documents from corpora such as C4 and Wikipedia; then we generate instructions for these documents via LLMs. This approach provides a cheaper and cleaner instruction-tuning dataset with natural output and one suitable for long text generation. Our models outperform 10x larger language models without instruction tuning on tasks such as story/recipe generation and long-form question answering. Moreover, LongForm models outperform prior instruction-tuned models such as FLAN-T5 and Alpaca by a large margin, and improve language understanding capabilities further. We publicly release our data and models: https://github.com/akoksal/LongForm.

comment: EMNLP 2024 Findings. This version extends the training with recent LLMs, evaluation with new metrics, and NLU tasks

♻ ☆ TurkishMMLU: Measuring Massive Multitask Language Understanding in Turkish EMNLP 2024

Arda Yüksel, Abdullatif Köksal, Lütfi Kerem Şenel, Anna Korhonen, Hinrich Schütze

Multiple choice question answering tasks evaluate the reasoning, comprehension, and mathematical abilities of Large Language Models (LLMs). While existing benchmarks employ automatic translation for multilingual evaluation, this approach is error-prone and potentially introduces culturally biased questions, especially in social sciences. We introduce the first multitask, multiple-choice Turkish QA benchmark, TurkishMMLU, to evaluate LLMs' understanding of the Turkish language. TurkishMMLU includes over 10,000 questions, covering 9 different subjects from Turkish high-school education curricula. These questions are written by curriculum experts, suitable for the high-school curricula in Turkey, covering subjects ranging from natural sciences and math questions to more culturally representative topics such as Turkish Literature and the history of the Turkish Republic. We evaluate over 20 LLMs, including multilingual open-source (e.g., Gemma, Llama, MT5), closed-source (GPT 4o, Claude, Gemini), and Turkish-adapted (e.g., Trendyol) models. We provide an extensive evaluation, including zero-shot and few-shot evaluation of LLMs, chain-of-thought reasoning, and question difficulty analysis along with model performance. We provide an in-depth analysis of the Turkish capabilities and limitations of current LLMs to provide insights for future LLMs for the Turkish language. We publicly release our code for the dataset and evaluation: https://github.com/ArdaYueksel/TurkishMMLU.

comment: EMNLP 2024 - Findings

♻ ☆ Generate-on-Graph: Treat LLM as both Agent and KG in Incomplete Knowledge Graph Question Answering EMNLP 2024

Yao Xu, Shizhu He, Jiabei Chen, Zihao Wang, Yangqiu Song, Hanghang Tong, Guang Liu, Kang Liu, Jun Zhao

To address the issues of insufficient knowledge and hallucination in Large Language Models (LLMs), numerous studies have explored integrating LLMs with Knowledge Graphs (KGs). However, these methods are typically evaluated on conventional Knowledge Graph Question Answering (KGQA) with complete KGs, where all factual triples required for each question are entirely covered by the given KG. In such cases, LLMs primarily act as an agent to find answer entities within the KG, rather than effectively integrating the internal knowledge of LLMs and external knowledge sources such as KGs. In fact, KGs are often incomplete to cover all the knowledge required to answer questions. To simulate these real-world scenarios and evaluate the ability of LLMs to integrate internal and external knowledge, we propose leveraging LLMs for QA under Incomplete Knowledge Graph (IKGQA), where the provided KG lacks some of the factual triples for each question, and construct corresponding datasets. To handle IKGQA, we propose a training-free method called Generate-on-Graph (GoG), which can generate new factual triples while exploring KGs. Specifically, GoG performs reasoning through a Thinking-Searching-Generating framework, which treats LLM as both Agent and KG in IKGQA. Experimental results on two datasets demonstrate that our GoG outperforms all previous methods.

comment: Accepted by EMNLP 2024 Main

♻ ☆ Quantifying Generalization Complexity for Large Language Models

Zhenting Qi, Hongyin Luo, Xuliang Huang, Zhuokai Zhao, Yibo Jiang, Xiangjun Fan, Himabindu Lakkaraju, James Glass

While large language models (LLMs) have shown exceptional capabilities in understanding complex queries and performing sophisticated tasks, their generalization abilities are often deeply entangled with memorization, necessitating more precise evaluation. To address this challenge, we introduce Scylla, a dynamic evaluation framework that quantitatively measures the generalization abilities of LLMs. Scylla disentangles generalization from memorization via assessing model performance on both in-distribution (ID) and out-of-distribution (OOD) data through 20 tasks across 5 levels of complexity. Through extensive experiments, we uncover a non-monotonic relationship between task complexity and the performance gap between ID and OOD data, which we term the generalization valley. Specifically, this phenomenon reveals a critical threshold - referred to as critical complexity - where reliance on non-generalizable behavior peaks, indicating the upper bound of LLMs' generalization capabilities. As model size increases, the critical complexity shifts toward higher levels of task complexity, suggesting that larger models can handle more complex reasoning tasks before over-relying on memorization. Leveraging Scylla and the concept of critical complexity, we benchmark 28LLMs including both open-sourced models such as LLaMA and Qwen families, and close-sourced models like Claude and GPT, providing a more robust evaluation and establishing a clearer understanding of LLMs' generalization capabilities.

♻ ☆ Ada-Instruct: Adapting Instruction Generators for Complex Reasoning

Wanyun Cui, Qianle Wang

Instructions augmentation is a crucial step for unleashing the full potential of large language models (LLMs) in downstream tasks. Existing Self-Instruct methods primarily simulate new instructions from a few initial instructions with in-context learning. However, our study identifies a critical flaw in this approach: even with GPT4o, Self-Instruct cannot generate complex instructions of length $\ge 100$, which is necessary in complex tasks such as code completion. To address this issue, our key insight is that fine-tuning open source LLMs with only ten examples can produce complex instructions that maintain distributional consistency for complex reasoning tasks. We introduce Ada-Instruct, an adaptive instruction generator developed through fine-tuning. We empirically validated Ada-Instruct's efficacy across different applications. The results highlight Ada-Instruct's capacity to generate long, intricate, and distributionally consistent instructions.

♻ ☆ In-Context Editing: Learning Knowledge from Self-Induced Distributions

Siyuan Qi, Bangcheng Yang, Kailin Jiang, Xiaobo Wang, Jiaqi Li, Yifan Zhong, Yaodong Yang, Zilong Zheng

In scenarios where language models must incorporate new information efficiently without extensive retraining, traditional fine-tuning methods are prone to overfitting, degraded generalization, and unnatural language generation. To address these limitations, we introduce Consistent In-Context Editing (ICE), a novel approach leveraging the model's in-context learning capability to optimize toward a contextual distribution rather than a one-hot target. ICE introduces a simple yet effective optimization framework for the model to internalize new knowledge by aligning its output distributions with and without additional context. This method enhances the robustness and effectiveness of gradient-based tuning methods, preventing overfitting and preserving the model's integrity. We analyze ICE across four critical aspects of knowledge editing: accuracy, locality, generalization, and linguistic quality, demonstrating its advantages. Experimental results confirm the effectiveness of ICE and demonstrate its potential for continual editing, ensuring that the integrity of the model is preserved while updating information.

♻ ☆ Fighting Randomness with Randomness: Mitigating Optimisation Instability of Fine-Tuning using Delayed Ensemble and Noisy Interpolation EMNLP'24

Branislav Pecher, Jan Cegin, Robert Belanec, Jakub Simko, Ivan Srba, Maria Bielikova

While fine-tuning of pre-trained language models generally helps to overcome the lack of labelled training samples, it also displays model performance instability. This instability mainly originates from randomness in initialisation or data shuffling. To address this, researchers either modify the training process or augment the available samples, which typically results in increased computational costs. We propose a new mitigation strategy, called Delayed Ensemble with Noisy Interpolation (DENI), that leverages the strengths of ensembling, noise regularisation and model interpolation, while retaining computational efficiency. We compare DENI with 9 representative mitigation strategies across 3 models, 4 tuning strategies and 7 text classification datasets. We show that: 1) DENI outperforms the best performing mitigation strategy (Ensemble), while using only a fraction of its cost; 2) the mitigation strategies are beneficial for parameter-efficient fine-tuning (PEFT) methods, outperforming full fine-tuning in specific cases; and 3) combining DENI with data augmentation often leads to even more effective instability mitigation.

comment: Accepted to the Findings of the EMNLP'24 Conference

♻ ☆ On Sensitivity of Learning with Limited Labelled Data to the Effects of Randomness: Impact of Interactions and Systematic Choices EMNLP'24

Branislav Pecher, Ivan Srba, Maria Bielikova

While learning with limited labelled data can improve performance when the labels are lacking, it is also sensitive to the effects of uncontrolled randomness introduced by so-called randomness factors (e.g., varying order of data). We propose a method to systematically investigate the effects of randomness factors while taking the interactions between them into consideration. To measure the true effects of an individual randomness factor, our method mitigates the effects of other factors and observes how the performance varies across multiple runs. Applying our method to multiple randomness factors across in-context learning and fine-tuning approaches on 7 representative text classification tasks and meta-learning on 3 tasks, we show that: 1) disregarding interactions between randomness factors in existing works caused inconsistent findings due to incorrect attribution of the effects of randomness factors, such as disproving the consistent sensitivity of in-context learning to sample order even with random sample selection; and 2) besides mutual interactions, the effects of randomness factors, especially sample order, are also dependent on more systematic choices unexplored in existing works, such as number of classes, samples per class or choice of prompt format.

comment: Accepted to the EMNLP'24 Main Conference

♻ ☆ 2D-TPE: Two-Dimensional Positional Encoding Enhances Table Understanding for Large Language Models

Jia-Nan Li, Jian Guan, Wei Wu, Zhengtao Yu, Rui Yan

Tables are ubiquitous across various domains for concisely representing structured information. Empowering large language models (LLMs) to reason over tabular data represents an actively explored direction. However, since typical LLMs only support one-dimensional~(1D) inputs, existing methods often flatten the two-dimensional~(2D) table structure into a sequence of tokens, which can severely disrupt the spatial relationships and result in an inevitable loss of vital contextual information. In this paper, we first empirically demonstrate the detrimental impact of such flattening operations on the performance of LLMs in capturing the spatial information of tables through two elaborate proxy tasks. Subsequently, we introduce a simple yet effective positional encoding method, termed ``2D-TPE'' (Two-Dimensional Table Positional Encoding), to address this challenge. 2D-TPE enables each attention head to dynamically select a permutation order of tokens within the context for attending to them, where each permutation represents a distinct traversal mode for the table, such as column-wise or row-wise traversal. 2D-TPE effectively mitigates the risk of losing essential spatial information while preserving computational efficiency, thus better preserving the table structure. Extensive experiments across five benchmarks demonstrate that 2D-TPE outperforms strong baselines, underscoring the importance of preserving the table structure for accurate table comprehension. Comprehensive analysis further reveals the substantially better scalability of 2D-TPE to large tables than baselines.

♻ ☆ TCSinger: Zero-Shot Singing Voice Synthesis with Style Transfer and Multi-Level Style Control EMNLP 2024

Yu Zhang, Ziyue Jiang, Ruiqi Li, Changhao Pan, Jinzheng He, Rongjie Huang, Chuxin Wang, Zhou Zhao

Zero-shot singing voice synthesis (SVS) with style transfer and style control aims to generate high-quality singing voices with unseen timbres and styles (including singing method, emotion, rhythm, technique, and pronunciation) from audio and text prompts. However, the multifaceted nature of singing styles poses a significant challenge for effective modeling, transfer, and control. Furthermore, current SVS models often fail to generate singing voices rich in stylistic nuances for unseen singers. To address these challenges, we introduce TCSinger, the first zero-shot SVS model for style transfer across cross-lingual speech and singing styles, along with multi-level style control. Specifically, TCSinger proposes three primary modules: 1) the clustering style encoder employs a clustering vector quantization model to stably condense style information into a compact latent space; 2) the Style and Duration Language Model (S\&D-LM) concurrently predicts style information and phoneme duration, which benefits both; 3) the style adaptive decoder uses a novel mel-style adaptive normalization method to generate singing voices with enhanced details. Experimental results show that TCSinger outperforms all baseline models in synthesis quality, singer similarity, and style controllability across various tasks, including zero-shot style transfer, multi-level style control, cross-lingual style transfer, and speech-to-singing style transfer. Singing voice samples can be accessed at https://tcsinger.github.io/.

comment: Accepted by EMNLP 2024

♻ ☆ Multi-FAct: Assessing Factuality of Multilingual LLMs using FActScore

Sheikh Shafayat, Eunsu Kim, Juhyun Oh, Alice Oh

Evaluating the factuality of long-form large language model (LLM)-generated text is an important challenge. Recently there has been a surge of interest in factuality evaluation for English, but little is known about the factuality evaluation of multilingual LLMs, specially when it comes to long-form generation. %This paper systematically evaluates multilingual LLMs' factual accuracy across languages and geographic regions. We introduce a simple pipeline for multilingual factuality evaluation, by applying FActScore (Min et al., 2023) for diverse languages. In addition to evaluating multilingual factual generation, we evaluate the factual accuracy of long-form text generation in topics that reflect regional diversity. We also examine the feasibility of running the FActScore pipeline using non-English Wikipedia and provide comprehensive guidelines on multilingual factual evaluation for regionally diverse topics.

♻ ☆ Latte: Latent Attention for Linear Time Transformers

Rares Dolga, Marius Cobzarenco, David Barber

♻ ☆ Twists, Humps, and Pebbles: Multilingual Speech Recognition Models Exhibit Gender Performance Gaps EMNLP 2024

Giuseppe Attanasio, Beatrice Savoldi, Dennis Fucci, Dirk Hovy

Current automatic speech recognition (ASR) models are designed to be used across many languages and tasks without substantial changes. However, this broad language coverage hides performance gaps within languages, for example, across genders. Our study systematically evaluates the performance of two widely used multilingual ASR models on three datasets, encompassing 19 languages from eight language families and two speaking conditions. Our findings reveal clear gender disparities, with the advantaged group varying across languages and models. Surprisingly, those gaps are not explained by acoustic or lexical properties. However, probing internal model states reveals a correlation with gendered performance gap. That is, the easier it is to distinguish speaker gender in a language using probes, the more the gap reduces, favoring female speakers. Our results show that gender disparities persist even in state-of-the-art models. Our findings have implications for the improvement of multilingual ASR systems, underscoring the importance of accessibility to training data and nuanced evaluation to predict and mitigate gender gaps. We release all code and artifacts at https://github.com/g8a9/multilingual-asr-gender-gap.

comment: Accepted at EMNLP 2024. Code and artifacts at https://github.com/g8a9/multilingual-asr-gender-gap

♻ ☆ Conversational Feedback in Scripted versus Spontaneous Dialogues: A Comparative Analysis

Ildikó Pilán, Laurent Prévot, Hendrik Buschmeier, Pierre Lison

Scripted dialogues such as movie and TV subtitles constitute a widespread source of training data for conversational NLP models. However, there are notable linguistic differences between these dialogues and spontaneous interactions, especially regarding the occurrence of communicative feedback such as backchannels, acknowledgments, or clarification requests. This paper presents a quantitative analysis of such feedback phenomena in both subtitles and spontaneous conversations. Based on conversational data spanning eight languages and multiple genres, we extract lexical statistics, classifications from a dialogue act tagger, expert annotations and labels derived from a fine-tuned Large Language Model (LLM). Our main empirical findings are that (1) communicative feedback is markedly less frequent in subtitles than in spontaneous dialogues and (2) subtitles contain a higher proportion of negative feedback. We also show that dialogues generated by standard LLMs lie much closer to scripted dialogues than spontaneous interactions in terms of communicative feedback.

comment: Updated version for SIGdial 2024

♻ ☆ miniCTX: Neural Theorem Proving with (Long-)Contexts

Jiewen Hu, Thomas Zhu, Sean Welleck

Real-world formal theorem proving often depends on a wealth of context, including definitions, lemmas, comments, file structure, and other information. We introduce miniCTX, which tests a model's ability to prove formal mathematical theorems that depend on new context that is not seen during training. miniCTX contains theorems sourced from real Lean projects and textbooks, each associated with a context that can span tens of thousands of tokens. Models are tasked with proving a theorem given access to code from the theorem's repository, which contains context that is needed for the proof. As a baseline for miniCTX, we tested fine-tuning and prompting methods that condition theorem proving on preceding context. Both approaches substantially outperform traditional methods that rely solely on state information. We found that this ability to use context is not captured by previous benchmarks such as miniF2F. Alongside miniCTX, we offer ntp-toolkit for automatically extracting and annotating theorem proving data, making it easy to add new projects into miniCTX to ensure that contexts are not seen during training. miniCTX offers a challenging and realistic evaluation of neural theorem provers.

♻ ☆ Compositional Hardness of Code in Large Language Models -- A Probabilistic Perspective

Yotam Wolf, Binyamin Rothberg, Dorin Shteyman, Amnon Shashua

A common practice in large language model (LLM) usage for complex analytical tasks such as code generation, is to sample a solution for the entire task within the model's context window. Previous works have shown that subtask decomposition within the model's context (chain of thought), is beneficial for solving such tasks. In this work, we point a limitation of LLMs' ability to perform several sub-tasks within the same context window - an in-context hardness of composition, pointing to an advantage for distributing a decomposed problem in a multi-agent system of LLMs. The hardness of composition is quantified by a generation complexity metric, i.e., the number of LLM generations required to sample at least one correct solution. We find a gap between the generation complexity of solving a compositional problem within the same context relative to distributing it among multiple agents, that increases exponentially with the solution's length. We prove our results theoretically and demonstrate them empirically.

♻ ☆ Evaluating Automatic Metrics with Incremental Machine Translation Systems

Guojun Wu, Shay B. Cohen, Rico Sennrich

We introduce a dataset comprising commercial machine translations, gathered weekly over six years across 12 translation directions. Since human A/B testing is commonly used, we assume commercial systems improve over time, which enables us to evaluate machine translation (MT) metrics based on their preference for more recent translations. Our study not only confirms several prior findings, such as the advantage of neural metrics over non-neural ones, but also explores the debated issue of how MT quality affects metric reliability--an investigation that smaller datasets in previous research could not sufficiently explore. Overall, our research demonstrates the dataset's value as a testbed for metric evaluation. We release our code at https://github.com/gjwubyron/Evo

♻ ☆ Graph Chain-of-Thought: Augmenting Large Language Models by Reasoning on Graphs

Bowen Jin, Chulin Xie, Jiawei Zhang, Kashob Kumar Roy, Yu Zhang, Zheng Li, Ruirui Li, Xianfeng Tang, Suhang Wang, Yu Meng, Jiawei Han

Large language models (LLMs), while exhibiting exceptional performance, suffer from hallucinations, especially on knowledge-intensive tasks. Existing works propose to augment LLMs with individual text units retrieved from external knowledge corpora to alleviate the issue. However, in many domains, texts are interconnected (e.g., academic papers in a bibliographic graph are linked by citations and co-authorships) which form a (text-attributed) graph. The knowledge in such graphs is encoded not only in single texts/nodes but also in their associated connections. To facilitate the research of augmenting LLMs with graphs, we manually construct a Graph Reasoning Benchmark dataset called GRBench, containing 1,740 questions that can be answered with the knowledge from 10 domain graphs. Then, we propose a simple and effective framework called Graph Chain-of-thought (Graph-CoT) to augment LLMs with graphs by encouraging LLMs to reason on the graph iteratively. Each Graph-CoT iteration consists of three sub-steps: LLM reasoning, LLM-graph interaction, and graph execution. We conduct systematic experiments with three LLM backbones on GRBench, where Graph-CoT outperforms the baselines consistently. The code is available at https://github.com/PeterGriffinJin/Graph-CoT.

comment: 21 pages. Code: https://github.com/PeterGriffinJin/Graph-CoT

♻ ☆ Distilling Instruction-following Abilities of Large Language Models with Task-aware Curriculum Planning

Yuanhao Yue, Chengyu Wang, Jun Huang, Peng Wang

Instruction tuning aims to align large language models (LLMs) with open-domain instructions and human-preferred responses. While several studies have explored autonomous approaches to distilling and annotating instructions from powerful proprietary LLMs, such as ChatGPT, they often neglect the impact of the distributions and characteristics of tasks, together with the varying difficulty of instructions in training sets. This oversight can lead to imbalanced knowledge capabilities and poor generalization powers of student LLMs. To address these challenges, we introduce Task-Aware Curriculum Planning for Instruction Refinement (TAPIR), a multi-round distillation framework that utilizes an oracle LLM to select instructions that are difficult for a student LLM to follow. To balance the student's capabilities, task distributions in training sets are adjusted with responses automatically refined according to their corresponding tasks. In addition, by incorporating curriculum planning, our approach systematically escalates the difficulty levels of tasks, progressively enhancing the student LLM's capabilities. We rigorously evaluate TAPIR using several widely recognized benchmarks (such as AlpacaEval 2.0, MT-Bench, etc.) and multiple student LLMs. Empirical results demonstrate that student LLMs, trained with our method and less training data, outperform larger instruction-tuned models and strong distillation baselines.

comment: emnlp 2024 findings

♻ ☆ A Systematic Survey and Critical Review on Evaluating Large Language Models: Challenges, Limitations, and Recommendations EMNLP 2024

Md Tahmid Rahman Laskar, Sawsan Alqahtani, M Saiful Bari, Mizanur Rahman, Mohammad Abdullah Matin Khan, Haidar Khan, Israt Jahan, Amran Bhuiyan, Chee Wei Tan, Md Rizwan Parvez, Enamul Hoque, Shafiq Joty, Jimmy Huang

Large Language Models (LLMs) have recently gained significant attention due to their remarkable capabilities in performing diverse tasks across various domains. However, a thorough evaluation of these models is crucial before deploying them in real-world applications to ensure they produce reliable performance. Despite the well-established importance of evaluating LLMs in the community, the complexity of the evaluation process has led to varied evaluation setups, causing inconsistencies in findings and interpretations. To address this, we systematically review the primary challenges and limitations causing these inconsistencies and unreliable evaluations in various steps of LLM evaluation. Based on our critical review, we present our perspectives and recommendations to ensure LLM evaluations are reproducible, reliable, and robust.

comment: Accepted at EMNLP 2024 (Main Conference)

♻ ☆ Large Language Models on Graphs: A Comprehensive Survey

Bowen Jin, Gang Liu, Chi Han, Meng Jiang, Heng Ji, Jiawei Han

Large language models (LLMs), such as GPT4 and LLaMA, are creating significant advancements in natural language processing, due to their strong text encoding/decoding ability and newly found emergent capability (e.g., reasoning). While LLMs are mainly designed to process pure texts, there are many real-world scenarios where text data is associated with rich structure information in the form of graphs (e.g., academic networks, and e-commerce networks) or scenarios where graph data is paired with rich textual information (e.g., molecules with descriptions). Besides, although LLMs have shown their pure text-based reasoning ability, it is underexplored whether such ability can be generalized to graphs (i.e., graph-based reasoning). In this paper, we provide a systematic review of scenarios and techniques related to large language models on graphs. We first summarize potential scenarios of adopting LLMs on graphs into three categories, namely pure graphs, text-attributed graphs, and text-paired graphs. We then discuss detailed techniques for utilizing LLMs on graphs, including LLM as Predictor, LLM as Encoder, and LLM as Aligner, and compare the advantages and disadvantages of different schools of models. Furthermore, we discuss the real-world applications of such methods and summarize open-source codes and benchmark datasets. Finally, we conclude with potential future research directions in this fast-growing field. The related source can be found at https://github.com/PeterGriffinJin/Awesome-Language-Model-on-Graphs.

comment: 25 pages

♻ ☆ Tradeoffs Between Alignment and Helpfulness in Language Models with Representation Engineering

Yotam Wolf, Noam Wies, Dorin Shteyman, Binyamin Rothberg, Yoav Levine, Amnon Shashua

Language model alignment has become an important component of AI safety, allowing safe interactions between humans and language models, by enhancing desired behaviors and inhibiting undesired ones. It is often done by tuning the model or inserting preset aligning prompts. Recently, representation engineering, a method which alters the model's behavior via changing its representations post-training, was shown to be effective in aligning LLMs (Zou et al., 2023a). Representation engineering yields gains in alignment oriented tasks such as resistance to adversarial attacks and reduction of social biases, but was also shown to cause a decrease in the ability of the model to perform basic tasks. In this paper we study the tradeoff between the increase in alignment and decrease in helpfulness of the model. We propose a theoretical framework which provides bounds for these two quantities, and demonstrate their relevance empirically. First, we find that under the conditions of our framework, alignment can be guaranteed with representation engineering, and at the same time that helpfulness is harmed in the process. Second, we show that helpfulness is harmed quadratically with the norm of the representation engineering vector, while the alignment increases linearly with it, indicating a regime in which it is efficient to use representation engineering. We validate our findings empirically, and chart the boundaries to the usefulness of representation engineering for alignment.

♻ ☆ Eliciting In-Context Learning in Vision-Language Models for Videos Through Curated Data Distributional Properties EMNLP 2024

Keunwoo Peter Yu, Zheyuan Zhang, Fengyuan Hu, Shane Storks, Joyce Chai

A major reason behind the recent success of large language models (LLMs) is their \textit{in-context learning} capability, which makes it possible to rapidly adapt them to downstream text-based tasks by prompting them with a small number of relevant demonstrations. While large vision-language models (VLMs) have recently been developed for tasks requiring both text and images, they largely lack in-context learning over visual information, especially in understanding and generating text about videos. In this work, we implement \textbf{E}mergent \textbf{I}n-context \textbf{Le}arning on \textbf{V}ideos (\eilev{}), a novel training paradigm that induces in-context learning over video and text by capturing key properties of pre-training data found by prior work to be essential for in-context learning in transformers. In our experiments, we show that \eilev-trained models outperform other off-the-shelf VLMs in few-shot video narration for novel, rare actions. Furthermore, we demonstrate that these key properties of bursty distributions, skewed marginal distributions, and dynamic meaning each contribute to varying degrees to VLMs' in-context learning capability in narrating procedural videos. Our results, analysis, and \eilev{}-trained models yield numerous insights about the emergence of in-context learning over video and text, creating a foundation for future work to optimize and scale VLMs for open-domain video understanding and reasoning. Our code and demo are available at \url{https://github.com/yukw777/EILEV}.

comment: 16 pages, LaTeX; Accepted to EMNLP 2024 Main

♻ ☆ Does Instruction Tuning Make LLMs More Consistent?

Constanza Fierro, Jiaang Li, Anders Søgaard

The purpose of instruction tuning is enabling zero-shot performance, but instruction tuning has also been shown to improve chain-of-thought reasoning and value alignment (Si et al., 2023). Here we consider the impact on $\textit{consistency}$, i.e., the sensitivity of language models to small perturbations in the input. We compare 10 instruction-tuned LLaMA models to the original LLaMA-7b model and show that almost across-the-board they become more consistent, both in terms of their representations and their predictions in zero-shot and downstream tasks. We explain these improvements through mechanistic analyses of factual recall.

comment: We need to run extra experiments to ensure some of the claims in the paper are fully correct

♻ ☆ RGD: Multi-LLM Based Agent Debugger via Refinement and Generation Guidance

Haolin Jin, Zechao Sun, Huaming Chen

Large Language Models (LLMs) have shown incredible potential in code generation tasks, and recent research in prompt engineering have enhanced LLMs' understanding of textual information. However, ensuring the accuracy of generated code often requires extensive testing and validation by programmers. While LLMs can typically generate code based on task descriptions, their accuracy remains limited, especially for complex tasks that require a deeper understanding of both the problem statement and the code generation process. This limitation is primarily due to the LLMs' need to simultaneously comprehend text and generate syntactically and semantically correct code, without having the capability to automatically refine the code. In real-world software development, programmers rarely produce flawless code in a single attempt based on the task description alone, they rely on iterative feedback and debugging to refine their programs. Inspired by this process, we introduce a novel architecture of LLM-based agents for code generation and automatic debugging: Refinement and Guidance Debugging (RGD). The RGD framework is a multi-LLM-based agent debugger that leverages three distinct LLM agents-Guide Agent, Debug Agent, and Feedback Agent. RGD decomposes the code generation task into multiple steps, ensuring a clearer workflow and enabling iterative code refinement based on self-reflection and feedback. Experimental results demonstrate that RGD exhibits remarkable code generation capabilities, achieving state-of-the-art performance with a 9.8% improvement on the HumanEval dataset and a 16.2% improvement on the MBPP dataset compared to the state-of-the-art approaches and traditional direct prompting approaches. We highlight the effectiveness of the RGD framework in enhancing LLMs' ability to generate and refine code autonomously.

♻ ☆ Synthetic continued pretraining

Zitong Yang, Neil Band, Shuangping Li, Emmanuel Candès, Tatsunori Hashimoto

Pretraining on large-scale, unstructured internet text enables language models to acquire a significant amount of world knowledge. However, this knowledge acquisition is data-inefficient--to learn a given fact, models must be trained on hundreds to thousands of diverse representations of it. This poses a challenge when adapting a pretrained model to a small corpus of domain-specific documents, where each fact may appear rarely or only once. We propose to bridge this gap with synthetic continued pretraining: using the small domain-specific corpus to synthesize a large corpus more amenable to learning, and then performing continued pretraining on the synthesized corpus. We instantiate this proposal with EntiGraph, a synthetic data augmentation algorithm that extracts salient entities from the source documents and then generates diverse text by drawing connections between the sampled entities. Synthetic continued pretraining with EntiGraph enables a language model to answer questions and follow generic instructions related to the source documents without access to them. If, instead, the source documents are available at inference time, we show that the knowledge acquired through our approach compounds with retrieval-augmented generation. To better understand these results, we build a simple mathematical model of EntiGraph, and show how synthetic data augmentation can "rearrange" knowledge to enable more data-efficient learning.

comment: Updated organization of experimental results and methods introduction. Released the dataset and model weights artifact

♻ ☆ The SIFo Benchmark: Investigating the Sequential Instruction Following Ability of Large Language Models EMNLP 2024

Xinyi Chen, Baohao Liao, Jirui Qi, Panagiotis Eustratiadis, Christof Monz, Arianna Bisazza, Maarten de Rijke

Following multiple instructions is a crucial ability for large language models (LLMs). Evaluating this ability comes with significant challenges: (i) limited coherence between multiple instructions, (ii) positional bias where the order of instructions affects model performance, and (iii) a lack of objectively verifiable tasks. To address these issues, we introduce a benchmark designed to evaluate models' abilities to follow multiple instructions through sequential instruction following (SIFo) tasks. In SIFo, the successful completion of multiple instructions is verifiable by examining only the final instruction. Our benchmark evaluates instruction following using four tasks (text modification, question answering, mathematics, and security rules), each assessing different aspects of sequential instruction following. Our evaluation of popular LLMs, both closed-source and open-source, shows that more recent and larger models significantly outperform their older and smaller counterparts on the SIFo tasks, validating the benchmark's effectiveness. All models struggle with following sequences of instructions, hinting at an important lack of robustness of today's language models.

comment: EMNLP 2024 Findings

♻ ☆ Multilingual Synopses of Movie Narratives: A Dataset for Vision-Language Story Understanding

Yidan Sun, Jianfei Yu, Boyang Li

Story video-text alignment, a core task in computational story understanding, aims to align video clips with corresponding sentences in their descriptions. However, progress on the task has been held back by the scarcity of manually annotated video-text correspondence and the heavy concentration on English narrations of Hollywood movies. To address these issues, in this paper, we construct a large-scale multilingual video story dataset named Multilingual Synopses of Movie Narratives (M-SYMON), containing 13,166 movie summary videos from 7 languages, as well as manual annotation of fine-grained video-text correspondences for 101.5 hours of video. Training on the human annotated data from SyMoN outperforms the SOTA methods by 15.7 and 16.2 percentage points on Clip Accuracy and Sentence IoU scores, respectively, demonstrating the effectiveness of the annotations. As benchmarks for future research, we create 6 baseline approaches with different multilingual training strategies, compare their performance in both intra-lingual and cross-lingual setups, exemplifying the challenges of multilingual video-text alignment. The dataset is released at: https://github.com/insundaycathy/M-SyMoN

comment: 17 pages, 8 figures

♻ ☆ The Potential and Challenges of Evaluating Attitudes, Opinions, and Values in Large Language Models EMNLP 2024

Bolei Ma, Xinpeng Wang, Tiancheng Hu, Anna-Carolina Haensch, Michael A. Hedderich, Barbara Plank, Frauke Kreuter

Recent advances in Large Language Models (LLMs) have sparked wide interest in validating and comprehending the human-like cognitive-behavioral traits LLMs may capture and convey. These cognitive-behavioral traits include typically Attitudes, Opinions, Values (AOVs). However, measuring AOVs embedded within LLMs remains opaque, and different evaluation methods may yield different results. This has led to a lack of clarity on how different studies are related to each other and how they can be interpreted. This paper aims to bridge this gap by providing a comprehensive overview of recent works on the evaluation of AOVs in LLMs. Moreover, we survey related approaches in different stages of the evaluation pipeline in these works. By doing so, we address the potential and challenges with respect to understanding the model, human-AI alignment, and downstream application in social sciences. Finally, we provide practical insights into evaluation methods, model enhancement, and interdisciplinary collaboration, thereby contributing to the evolving landscape of evaluating AOVs in LLMs.

comment: EMNLP 2024 Findings

♻ ☆ Language models and brains align due to more than next-word prediction and word-level information EMNLP 2024

Gabriele Merlin, Mariya Toneva

Pretrained language models have been shown to significantly predict brain recordings of people comprehending language. Recent work suggests that the prediction of the next word is a key mechanism that contributes to this alignment. What is not yet understood is whether prediction of the next word is necessary for this observed alignment or simply sufficient, and whether there are other shared mechanisms or information that are similarly important. In this work, we take a step towards understanding the reasons for brain alignment via two simple perturbations in popular pretrained language models. These perturbations help us design contrasts that can control for different types of information. By contrasting the brain alignment of these differently perturbed models, we show that improvements in alignment with brain recordings are due to more than improvements in next-word prediction and word-level information.

comment: Accepted to EMNLP 2024

♻ ☆ Lexically Grounded Subword Segmentation EMNLP

Jindřich Libovický, Jindřich Helcl

We present three innovations in tokenization and subword segmentation. First, we propose to use unsupervised morphological analysis with Morfessor as pre-tokenization. Second, we present an algebraic method for obtaining subword embeddings grounded in a word embedding space. Based on that, we design a novel subword segmentation algorithm that uses the embeddings, ensuring that the procedure considers lexical meaning. Third, we introduce an efficient segmentation algorithm based on a subword bigram model that can be initialized with the lexically aware segmentation method to avoid using Morfessor and large embedding tables at inference time. We evaluate the proposed approaches using two intrinsic metrics and measure their performance on two downstream tasks: part-of-speech tagging and machine translation. Our experiments show significant improvements in the morphological plausibility of the segmentation when evaluated using segmentation precision on morpheme boundaries and improved R\'enyi efficiency in 8 languages. Although the proposed tokenization methods do not have a large impact on automatic translation quality, we observe consistent performance gains in the arguably more morphological task of part-of-speech tagging.

comment: Camera-ready, EMNLP Main conf

♻ ☆ Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation EMNLP 2024

Jirui Qi, Gabriele Sarti, Raquel Fernández, Arianna Bisazza

Ensuring the verifiability of model answers is a fundamental challenge for retrieval-augmented generation (RAG) in the question answering (QA) domain. Recently, self-citation prompting was proposed to make large language models (LLMs) generate citations to supporting documents along with their answers. However, self-citing LLMs often struggle to match the required format, refer to non-existent sources, and fail to faithfully reflect LLMs' context usage throughout the generation. In this work, we present MIRAGE --Model Internals-based RAG Explanations -- a plug-and-play approach using model internals for faithful answer attribution in RAG applications. MIRAGE detects context-sensitive answer tokens and pairs them with retrieved documents contributing to their prediction via saliency methods. We evaluate our proposed approach on a multilingual extractive QA dataset, finding high agreement with human answer attribution. On open-ended QA, MIRAGE achieves citation quality and efficiency comparable to self-citation while also allowing for a finer-grained control of attribution parameters. Our qualitative evaluation highlights the faithfulness of MIRAGE's attributions and underscores the promising application of model internals for RAG answer attribution.

comment: Accepted by EMNLP 2024 Main Conference. Code and data released at https://github.com/Betswish/MIRAGE

♻ ☆ LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via a Hybrid Architecture

Xidong Wang, Dingjie Song, Shunian Chen, Chen Zhang, Benyou Wang

Expanding the long-context capabilities of Multi-modal Large Language Models~(MLLMs) is crucial for video understanding, high-resolution image understanding, and multi-modal agents. This involves a series of systematic optimizations, including model architecture, data construction and training strategy, particularly addressing challenges such as \textit{degraded performance with more images} and \textit{high computational costs}. In this paper, we adapt the model architecture to a hybrid of Mamba and Transformer blocks, approach data construction with both temporal and spatial dependencies among multiple images and employ a progressive training strategy. The released model \textbf{LongLLaVA}~(\textbf{Long}-Context \textbf{L}arge \textbf{L}anguage \textbf{a}nd \textbf{V}ision \textbf{A}ssistant) is the first hybrid MLLM, which achieved a better balance between efficiency and effectiveness. LongLLaVA not only achieves competitive results across various benchmarks, but also maintains high throughput and low memory consumption. Especially, it could process nearly a thousand images on a single A100 80GB GPU, showing promising application prospects for a wide range of tasks.

comment: 20 pages, 9 figures, 9 tables

♻ ☆ PromptWizard: Task-Aware Prompt Optimization Framework

Eshaan Agarwal, Joykirat Singh, Vivek Dani, Raghav Magazine, Tanuja Ganu, Akshay Nambi

Large language models (LLMs) have transformed AI across diverse domains, with prompting being central to their success in guiding model outputs. However, manual prompt engineering is both labor-intensive and domain-specific, necessitating the need for automated solutions. We introduce PromptWizard, a novel, fully automated framework for discrete prompt optimization, utilizing a self-evolving, self-adapting mechanism. Through a feedback-driven critique and synthesis process, PromptWizard achieves an effective balance between exploration and exploitation, iteratively refining both prompt instructions and in-context examples to generate human-readable, task-specific prompts. This guided approach systematically improves prompt quality, resulting in superior performance across 45 tasks. PromptWizard excels even with limited training data, smaller LLMs, and various LLM architectures. Additionally, our cost analysis reveals a substantial reduction in API calls, token usage, and overall cost, demonstrating PromptWizard's efficiency, scalability, and advantages over existing prompt optimization strategies.

♻ ☆ On the Adversarial Vulnerability of Pairwise Evaluation Using Large Language Models

Hawon Jeong, ChaeHun Park, Jimin Hong, Jaegul Choo

Pairwise evaluation using large language models (LLMs) is widely adopted for evaluating generated outputs. However, the reliability of LLM evaluators is often compromised by their biased preferences, such as favoring verbosity and an authoritative tone. In this work, we find that the evaluation setup itself can significantly amplify these biases, where pairwise evaluators exhibit more undesirable tendencies than pointwise evaluators. Our analysis further reveals that even when pairwise evaluators make incorrect judgments, they can still accurately identify shortcomings in low-quality outputs. As a simple remedy, we also propose incorporating pointwise reasoning into pairwise evaluation. Experimental results show that our method improves the performance of pairwise evaluators on adversarial samples across various models. We hope our findings encourage further exploration into the reliability of LLM evaluators.

♻ ☆ Generalists vs. Specialists: Evaluating Large Language Models for Urdu

Samee Arif, Abdul Hameed Azeemi, Agha Ali Raza, Awais Athar

In this paper, we compare general-purpose models, GPT-4-Turbo and Llama-3-8b, with special-purpose models--XLM-Roberta-large, mT5-large, and Llama-3-8b--that have been fine-tuned on specific tasks. We focus on seven classification and seven generation tasks to evaluate the performance of these models on Urdu language. Urdu has 70 million native speakers, yet it remains underrepresented in Natural Language Processing (NLP). Despite the frequent advancements in Large Language Models (LLMs), their performance in low-resource languages, including Urdu, still needs to be explored. We also conduct a human evaluation for the generation tasks and compare the results with the evaluations performed by GPT-4-Turbo, Llama-3-8b and Claude 3.5 Sonnet. We find that special-purpose models consistently outperform general-purpose models across various tasks. We also find that the evaluation done by GPT-4-Turbo for generation tasks aligns more closely with human evaluation compared to the evaluation the evaluation done by Llama-3-8b. This paper contributes to the NLP community by providing insights into the effectiveness of general and specific-purpose LLMs for low-resource languages.

♻ ☆ Efficient Temporal Extrapolation of Multimodal Large Language Models with Temporal Grounding Bridge EMNLP 2024

Yuxuan Wang, Yueqian Wang, Pengfei Wu, Jianxin Liang, Dongyan Zhao, Yang Liu, Zilong Zheng

Despite progress in multimodal large language models (MLLMs), the challenge of interpreting long-form videos in response to linguistic queries persists, largely due to the inefficiency in temporal grounding and limited pre-trained context window size. In this work, we introduce Temporal Grounding Bridge (TGB), a novel framework that bootstraps MLLMs with advanced temporal grounding capabilities and broadens their contextual scope. Our framework significantly enhances the temporal capabilities of current MLLMs through three key innovations: an efficient multi-span temporal grounding algorithm applied to low-dimension temporal features projected from flow; a multimodal length extrapolation training paradigm that utilizes low-dimension temporal features to extend the training context window size; and a bootstrapping framework that bridges our model with pluggable MLLMs without requiring annotation. We validate TGB across seven video benchmarks and demonstrate substantial performance improvements compared with prior MLLMs. Notably, our model, initially trained on sequences of four frames, effectively handles sequences up to 16 longer without sacrificing performance, highlighting its scalability and effectiveness in real-world applications. Our code is publicly available at https://github.com/bigai-nlco/VideoTGB

comment: To appear at EMNLP 2024

♻ ☆ WaveletGPT: Wavelets Meet Large Language Models

Prateek Verma

Large Language Models (LLMs) have ushered in a new wave of artificial intelligence advancements impacting every scientific field and discipline. They are trained on a simple objective: to predict the next token given the previous context. We live in a world where most of the data around us, e.g., text, audio, and music, has a multi-scale structure associated with it. This paper infuses LLMs with traditional signal processing ideas, namely wavelets, during pre-training to take advantage of the structure. Without adding \textbf{any extra parameters} to a GPT-style LLM architecture, we achieve the same pre-training performance almost twice as fast in text, raw audio, and symbolic music. This is achieved by imposing a structure on intermediate embeddings. When trained for the same number of training steps, we achieve significant gains in performance, which is comparable to pre-training a larger neural architecture. Our architecture allows every next token prediction access to intermediate embeddings at different temporal resolutions in every Transformer decoder block. This work will hopefully pave the way for incorporating multi-rate signal processing ideas into traditional LLM pre-training. Further, we showcase pushing model performance by improving internal structure instead of just going after scale.

comment: 16 pages, 4 figures

♻ ☆ A Comprehensive Survey of Hallucination in Large Language, Image, Video and Audio Foundation Models EMNLP 2024

Pranab Sahoo, Prabhash Meharia, Akash Ghosh, Sriparna Saha, Vinija Jain, Aman Chadha

The rapid advancement of foundation models (FMs) across language, image, audio, and video domains has shown remarkable capabilities in diverse tasks. However, the proliferation of FMs brings forth a critical challenge: the potential to generate hallucinated outputs, particularly in high-stakes applications. The tendency of foundation models to produce hallucinated content arguably represents the biggest hindrance to their widespread adoption in real-world scenarios, especially in domains where reliability and accuracy are paramount. This survey paper presents a comprehensive overview of recent developments that aim to identify and mitigate the problem of hallucination in FMs, spanning text, image, video, and audio modalities. By synthesizing recent advancements in detecting and mitigating hallucination across various modalities, the paper aims to provide valuable insights for researchers, developers, and practitioners. Essentially, it establishes a clear framework encompassing definition, taxonomy, and detection strategies for addressing hallucination in multimodal foundation models, laying the foundation for future research in this pivotal area.

comment: EMNLP 2024 Findings

♻ ☆ PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling

Zefan Cai, Yichi Zhang, Bofei Gao, Yuliang Liu, Tianyu Liu, Keming Lu, Wayne Xiong, Yue Dong, Baobao Chang, Junjie Hu, Wen Xiao

In this study, we investigate whether attention-based information flow inside large language models (LLMs) is aggregated through noticeable patterns for long context processing. Our observations reveal that LLMs aggregate information through Pyramidal Information Funneling where attention is scattering widely in lower layers, progressively consolidating within specific contexts, and ultimately focusing on critical tokens (a.k.a massive activation or attention sink) in higher layers. Motivated by these insights, we developed PyramidKV, a novel and effective KV cache compression method. This approach dynamically adjusts the KV cache size across different layers, allocating more cache in lower layers and less in higher ones, diverging from traditional methods that maintain a uniform KV cache size. Our experimental evaluations, utilizing the LongBench benchmark, show that PyramidKV matches the performance of models with a full KV cache while retaining only 12% of the KV cache, thus significantly reducing memory usage. In scenarios emphasizing memory efficiency, where only 0.7% of the KV cache is maintained, PyramidKV surpasses other KV cache compression techniques, achieving up to a 20.5 absolute accuracy improvement on TREC dataset. In the Needle-in-a-Haystack experiment, PyramidKV outperforms competing methods in maintaining long-context comprehension in LLMs; notably, retaining just 128 KV cache entries enables the LLAMA-3-70B model to achieve 100% Acc. performance, matching that of a full KV cache.

♻ ☆ Self-Constructed Context Decompilation with Fined-grained Alignment Enhancement EMNLP 2024

Yunlong Feng, Dechuan Teng, Yang Xu, Honglin Mu, Xiao Xu, Libo Qin, Qingfu Zhu, Wanxiang Che

Decompilation transforms compiled code back into a high-level programming language for analysis when source code is unavailable. Previous work has primarily focused on enhancing decompilation performance by increasing the scale of model parameters or training data for pre-training. Based on the characteristics of the decompilation task, we propose two methods: (1) Without fine-tuning, the Self-Constructed Context Decompilation (sc$^2$dec) method recompiles the LLM's decompilation results to construct pairs for in-context learning, helping the model improve decompilation performance. (2) Fine-grained Alignment Enhancement (FAE), which meticulously aligns assembly code with source code at the statement level by leveraging debugging information, is employed during the fine-tuning phase to achieve further improvements in decompilation. By integrating these two methods, we achieved a Re-Executability performance improvement of approximately 3.90% on the Decompile-Eval benchmark, establishing a new state-of-the-art performance of 52.41%. The code, data, and models are available at https://github.com/AlongWY/sccdec.

comment: EMNLP 2024 Findings

♻ ☆ Exploring language relations through syntactic distances and geographic proximity

Juan De Gregorio, Raúl Toral, David Sánchez

Languages are grouped into families that share common linguistic traits. While this approach has been successful in understanding genetic relations between diverse languages, more analyses are needed to accurately quantify their relatedness, especially in less studied linguistic levels such as syntax. Here, we explore linguistic distances using series of parts of speech (POS) extracted from the Universal Dependencies dataset. Within an information-theoretic framework, we show that employing POS trigrams maximizes the possibility of capturing syntactic variations while being at the same time compatible with the amount of available data. Linguistic connections are then established by assessing pairwise distances based on the POS distributions. Intriguingly, our analysis reveals definite clusters that correspond to well known language families and groups, with exceptions explained by distinct morphological typologies. Furthermore, we obtain a significant correlation between language similarity and geographic distance, which underscores the influence of spatial proximity on language kinships.

comment: 39 pages

♻ ☆ A Systematic Analysis of Large Language Models as Soft Reasoners: The Case of Syllogistic Inferences

Leonardo Bertolazzi, Albert Gatt, Raffaella Bernardi

The reasoning abilities of Large Language Models (LLMs) are becoming a central focus of study in NLP. In this paper, we consider the case of syllogistic reasoning, an area of deductive reasoning studied extensively in logic and cognitive psychology. Previous research has shown that pre-trained LLMs exhibit reasoning biases, such as $\textit{content effects}$, avoid answering that $\textit{no conclusion follows}$, display human-like difficulties, and struggle with multi-step reasoning. We contribute to this research line by systematically investigating the effects of chain-of-thought reasoning, in-context learning (ICL), and supervised fine-tuning (SFT) on syllogistic reasoning, considering syllogisms with conclusions that support or violate world knowledge, as well as ones with multiple premises. Crucially, we go beyond the standard focus on accuracy, with an in-depth analysis of the conclusions generated by the models. Our results suggest that the behavior of pre-trained LLMs can be explained by heuristics studied in cognitive science and that both ICL and SFT improve model performance on valid inferences, although only the latter mitigates most reasoning biases without harming model consistency.

♻ ☆ Optimized Speculative Sampling for GPU Hardware Accelerators EMNLP 2024

Dominik Wagner, Seanie Lee, Ilja Baumann, Philipp Seeberger, Korbinian Riedhammer, Tobias Bocklet

In this work, we optimize speculative sampling for parallel hardware accelerators to improve sampling speed. We notice that substantial portions of the intermediate matrices necessary for speculative sampling can be computed concurrently. This allows us to distribute the workload across multiple GPU threads, enabling simultaneous operations on matrix segments within thread blocks. This results in profiling time improvements ranging from 6% to 13% relative to the baseline implementation, without compromising accuracy. To further accelerate speculative sampling, probability distributions parameterized by softmax are approximated by sigmoid. This approximation approach results in significantly greater relative improvements in profiling time, ranging from 37% to 94%, with a minor decline in accuracy. We conduct extensive experiments on both automatic speech recognition and summarization tasks to validate the effectiveness of our optimization methods.

comment: Accepted at EMNLP 2024

♻ ☆ StablePT: Towards Stable Prompting for Few-shot Learning via Input Separation EMNLP 2024

Xiaoming Liu, Chen Liu, Zhaohan Zhang, Chengzhengxu Li, Longtian Wang, Yu Lan, Chao Shen

Large language models have shown their ability to become effective few-shot learners with prompting, revolutionizing the paradigm of learning with data scarcity. However, this approach largely depends on the quality of prompt initialization, and always exhibits large variability among different runs. Such property makes prompt tuning highly unreliable and vulnerable to poorly constructed prompts, which limits its extension to more real-world applications. To tackle this issue, we propose to treat the hard prompt and soft prompt as separate inputs to mitigate noise brought by the prompt initialization. Furthermore, we optimize soft prompts with contrastive learning for utilizing class-aware information in the training process to maintain model performance. Experimental results demonstrate that \sysname outperforms state-of-the-art methods by 6.97% in accuracy and reduces the standard deviation by 1.92 on average. Furthermore, extensive experiments underscore its robustness and stability across 8 datasets covering various tasks. Codes are available at https://github.com/lccc0528/Stable/tree/main.

comment: EMNLP 2024 Findings

Computer Vision and Pattern Recognition 164

☆ Flash-Splat: 3D Reflection Removal with Flash Cues and Gaussian Splats

Mingyang Xie, Haoming Cai, Sachin Shah, Yiran Xu, Brandon Y. Feng, Jia-Bin Huang, Christopher A. Metzler

We introduce a simple yet effective approach for separating transmitted and reflected light. Our key insight is that the powerful novel view synthesis capabilities provided by modern inverse rendering methods (e.g.,~3D Gaussian splatting) allow one to perform flash/no-flash reflection separation using unpaired measurements -- this relaxation dramatically simplifies image acquisition over conventional paired flash/no-flash reflection separation methods. Through extensive real-world experiments, we demonstrate our method, Flash-Splat, accurately reconstructs both transmitted and reflected scenes in 3D. Our method outperforms existing 3D reflection separation methods, which do not leverage illumination control, by a large margin. Our project webpage is at https://flash-splat.github.io/.

☆ Vinoground: Scrutinizing LMMs over Dense Temporal Reasoning with Short Videos

Jianrui Zhang, Mu Cai, Yong Jae Lee

comment: Project Page: https://vinoground.github.io

☆ Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations

Nick Jiang, Anish Kachinthaya, Suzie Petryk, Yossi Gandelsman

We investigate the internal representations of vision-language models (VLMs) to address hallucinations, a persistent challenge despite advances in model size and training. We project VLMs' internal image representations to their language vocabulary and observe more confident output probabilities on real objects than hallucinated objects. We additionally use these output probabilities to spatially localize real objects. Building on this approach, we introduce a knowledge erasure algorithm that removes hallucinations by linearly orthogonalizing image features with respect to hallucinated object features. We show that targeted edits to a model's latent representations can reduce hallucinations by up to 25.7% on the COCO2014 dataset while preserving performance. Our findings demonstrate how a deeper understanding of VLMs' latent representations can enhance reliability and enable novel capabilities, such as zero-shot segmentation.

comment: Project page and code: http://anishk23733.github.io/vl-interp/

☆ FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Models

Zhipei Xu, Xuanyu Zhang, Runyi Li, Zecheng Tang, Qing Huang, Jian Zhang

The rapid development of generative AI is a double-edged sword, which not only facilitates content creation but also makes image manipulation easier and more difficult to detect. Although current image forgery detection and localization (IFDL) methods are generally effective, they tend to face two challenges: \textbf{1)} black-box nature with unknown detection principle, \textbf{2)} limited generalization across diverse tampering methods (e.g., Photoshop, DeepFake, AIGC-Editing). To address these issues, we propose the explainable IFDL task and design FakeShield, a multi-modal framework capable of evaluating image authenticity, generating tampered region masks, and providing a judgment basis based on pixel-level and image-level tampering clues. Additionally, we leverage GPT-4o to enhance existing IFDL datasets, creating the Multi-Modal Tamper Description dataSet (MMTD-Set) for training FakeShield's tampering analysis capabilities. Meanwhile, we incorporate a Domain Tag-guided Explainable Forgery Detection Module (DTE-FDM) and a Multi-modal Forgery Localization Module (MFLM) to address various types of tamper detection interpretation and achieve forgery localization guided by detailed textual descriptions. Extensive experiments demonstrate that FakeShield effectively detects and localizes various tampering techniques, offering an explainable and superior solution compared to previous IFDL methods.

☆ Loong: Generating Minute-level Long Videos with Autoregressive Language Models

Yuqing Wang, Tianwei Xiong, Daquan Zhou, Zhijie Lin, Yang Zhao, Bingyi Kang, Jiashi Feng, Xihui Liu

It is desirable but challenging to generate content-rich long videos in the scale of minutes. Autoregressive large language models (LLMs) have achieved great success in generating coherent and long sequences of tokens in the domain of natural language processing, while the exploration of autoregressive LLMs for video generation is limited to generating short videos of several seconds. In this work, we conduct a deep analysis of the challenges that prevent autoregressive LLM-based video generators from generating long videos. Based on the observations and analysis, we propose Loong, a new autoregressive LLM-based video generator that can generate minute-long videos. Specifically, we model the text tokens and video tokens as a unified sequence for autoregressive LLMs and train the model from scratch. We propose progressive short-to-long training with a loss re-weighting scheme to mitigate the loss imbalance problem for long video training. We further investigate inference strategies, including video token re-encoding and sampling strategies, to diminish error accumulation during inference. Our proposed Loong can be trained on 10-second videos and be extended to generate minute-level long videos conditioned on text prompts, as demonstrated by the results. More samples are available at: https://epiphqny.github.io/Loong-video.

comment: Project page: https://epiphqny.github.io/Loong-video/

☆ Contrastive Localized Language-Image Pre-Training

Hong-You Chen, Zhengfeng Lai, Haotian Zhang, Xinze Wang, Marcin Eichner, Keen You, Meng Cao, Bowen Zhang, Yinfei Yang, Zhe Gan

Contrastive Language-Image Pre-training (CLIP) has been a celebrated method for training vision encoders to generate image/text representations facilitating various applications. Recently, CLIP has been widely adopted as the vision backbone of multimodal large language models (MLLMs) to connect image inputs for language interactions. The success of CLIP as a vision-language foundation model relies on aligning web-crawled noisy text annotations at image levels. Nevertheless, such criteria may become insufficient for downstream tasks in need of fine-grained vision representations, especially when region-level understanding is demanding for MLLMs. In this paper, we improve the localization capability of CLIP with several advances. We propose a pre-training method called Contrastive Localized Language-Image Pre-training (CLOC) by complementing CLIP with region-text contrastive loss and modules. We formulate a new concept, promptable embeddings, of which the encoder produces image embeddings easy to transform into region representations given spatial hints. To support large-scale pre-training, we design a visually-enriched and spatially-localized captioning framework to effectively generate region-text pseudo-labels at scale. By scaling up to billions of annotated images, CLOC enables high-quality regional embeddings for image region recognition and retrieval tasks, and can be a drop-in replacement of CLIP to enhance MLLMs, especially on referring and grounding tasks.

comment: Preprint

☆ Revisit Large-Scale Image-Caption Data in Pre-training Multimodal Foundation Models

Zhengfeng Lai, Vasileios Saveris, Chen Chen, Hong-You Chen, Haotian Zhang, Bowen Zhang, Juan Lao Tebar, Wenze Hu, Zhe Gan, Peter Grasch, Meng Cao, Yinfei Yang

Recent advancements in multimodal models highlight the value of rewritten captions for improving performance, yet key challenges remain. For example, while synthetic captions often provide superior quality and image-text alignment, it is not clear whether they can fully replace AltTexts: the role of synthetic captions and their interaction with original web-crawled AltTexts in pre-training is still not well understood. Moreover, different multimodal foundation models may have unique preferences for specific caption formats, but efforts to identify the optimal captions for each model remain limited. In this work, we propose a novel, controllable, and scalable captioning pipeline designed to generate diverse caption formats tailored to various multimodal models. By examining Short Synthetic Captions (SSC) towards Dense Synthetic Captions (DSC+) as case studies, we systematically explore their effects and interactions with AltTexts across models such as CLIP, multimodal LLMs, and diffusion models. Our findings reveal that a hybrid approach that keeps both synthetic captions and AltTexts can outperform the use of synthetic captions alone, improving both alignment and performance, with each model demonstrating preferences for particular caption formats. This comprehensive analysis provides valuable insights into optimizing captioning strategies, thereby advancing the pre-training of multimodal foundation models.

comment: CV/ML

☆ DivScene: Benchmarking LVLMs for Object Navigation with Diverse Scenes and Objects

Zhaowei Wang, Hongming Zhang, Tianqing Fang, Ye Tian, Yue Yang, Kaixin Ma, Xiaoman Pan, Yangqiu Song, Dong Yu

comment: Work in Progress

☆ Curvature Diversity-Driven Deformation and Domain Alignment for Point Cloud

Mengxi Wu, Hao Huang, Yi Fang, Mohammad Rostami

Unsupervised Domain Adaptation (UDA) is crucial for reducing the need for extensive manual data annotation when training deep networks on point cloud data. A significant challenge of UDA lies in effectively bridging the domain gap. To tackle this challenge, we propose \textbf{C}urvature \textbf{D}iversity-Driven \textbf{N}uclear-Norm Wasserstein \textbf{D}omain Alignment (CDND). Our approach first introduces a \textit{\textbf{Curv}ature Diversity-driven Deformation \textbf{Rec}onstruction (CurvRec)} task, which effectively mitigates the gap between the source and target domains by enabling the model to extract salient features from semantically rich regions of a given point cloud. We then propose \textit{\textbf{D}eformation-based \textbf{N}uclear-norm \textbf{W}asserstein \textbf{D}iscrepancy (D-NWD)}, which applies the Nuclear-norm Wasserstein Discrepancy to both \textit{deformed and original} data samples to align the source and target domains. Furthermore, we contribute a theoretical justification for the effectiveness of D-NWD in distribution alignment and demonstrate that it is \textit{generic} enough to be applied to \textbf{any} deformations. To validate our method, we conduct extensive experiments on two public domain adaptation datasets for point cloud classification and segmentation tasks. Empirical experiment results show that our CDND achieves state-of-the-art performance by a noticeable margin over existing approaches.

☆ AlzhiNet: Traversing from 2DCNN to 3DCNN, Towards Early Detection and Diagnosis of Alzheimer's Disease

Romoke Grace Akindele, Samuel Adebayo, Paul Shekonya Kanda, Ming Yu

Alzheimer's disease (AD) is a progressive neurodegenerative disorder with increasing prevalence among the aging population, necessitating early and accurate diagnosis for effective disease management. In this study, we present a novel hybrid deep learning framework that integrates both 2D Convolutional Neural Networks (2D-CNN) and 3D Convolutional Neural Networks (3D-CNN), along with a custom loss function and volumetric data augmentation, to enhance feature extraction and improve classification performance in AD diagnosis. According to extensive experiments, AlzhiNet outperforms standalone 2D and 3D models, highlighting the importance of combining these complementary representations of data. The depth and quality of 3D volumes derived from the augmented 2D slices also significantly influence the model's performance. The results indicate that carefully selecting weighting factors in hybrid predictions is imperative for achieving optimal results. Our framework has been validated on the Magnetic Resonance Imaging (MRI) from Kaggle and MIRIAD datasets, obtaining accuracies of 98.9% and 99.99%, respectively, with an AUC of 100%. Furthermore, AlzhiNet was studied under a variety of perturbation scenarios on the Alzheimer's Kaggle dataset, including Gaussian noise, brightness, contrast, salt and pepper noise, color jitter, and occlusion. The results obtained show that AlzhiNet is more robust to perturbations than ResNet-18, making it an excellent choice for real-world applications. This approach represents a promising advancement in the early diagnosis and treatment planning for Alzheimer's disease.

☆ Video Instruction Tuning With Synthetic Data

Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, Chunyuan Li

comment: Project page: https://llava-vl.github.io/blog/2024-09-30-llava-video/

☆ LLaVA-Critic: Learning to Evaluate Multimodal Models

Tianyi Xiong, Xiyao Wang, Dong Guo, Qinghao Ye, Haoqi Fan, Quanquan Gu, Heng Huang, Chunyuan Li

comment: Project Page: https://llava-vl.github.io/blog/2024-10-03-llava-critic

☆ SteerDiff: Steering towards Safe Text-to-Image Diffusion Models

Hongxiang Zhang, Yifeng He, Hao Chen

Text-to-image (T2I) diffusion models have drawn attention for their ability to generate high-quality images with precise text alignment. However, these models can also be misused to produce inappropriate content. Existing safety measures, which typically rely on text classifiers or ControlNet-like approaches, are often insufficient. Traditional text classifiers rely on large-scale labeled datasets and can be easily bypassed by rephrasing. As diffusion models continue to scale, fine-tuning these safeguards becomes increasingly challenging and lacks flexibility. Recent red-teaming attack researches further underscore the need for a new paradigm to prevent the generation of inappropriate content. In this paper, we introduce SteerDiff, a lightweight adaptor module designed to act as an intermediary between user input and the diffusion model, ensuring that generated images adhere to ethical and safety standards with little to no impact on usability. SteerDiff identifies and manipulates inappropriate concepts within the text embedding space to guide the model away from harmful outputs. We conduct extensive experiments across various concept unlearning tasks to evaluate the effectiveness of our approach. Furthermore, we benchmark SteerDiff against multiple red-teaming strategies to assess its robustness. Finally, we explore the potential of SteerDiff for concept forgetting tasks, demonstrating its versatility in text-conditioned image generation.

☆ ControlAR: Controllable Image Generation with Autoregressive Models

Zongming Li, Tianheng Cheng, Shoufa Chen, Peize Sun, Haocheng Shen, Longjin Ran, Xiaoxin Chen, Wenyu Liu, Xinggang Wang

Autoregressive (AR) models have reformulated image generation as next-token prediction, demonstrating remarkable potential and emerging as strong competitors to diffusion models. However, control-to-image generation, akin to ControlNet, remains largely unexplored within AR models. Although a natural approach, inspired by advancements in Large Language Models, is to tokenize control images into tokens and prefill them into the autoregressive model before decoding image tokens, it still falls short in generation quality compared to ControlNet and suffers from inefficiency. To this end, we introduce ControlAR, an efficient and effective framework for integrating spatial controls into autoregressive image generation models. Firstly, we explore control encoding for AR models and propose a lightweight control encoder to transform spatial inputs (e.g., canny edges or depth maps) into control tokens. Then ControlAR exploits the conditional decoding method to generate the next image token conditioned on the per-token fusion between control and image tokens, similar to positional encodings. Compared to prefilling tokens, using conditional decoding significantly strengthens the control capability of AR models but also maintains the model's efficiency. Furthermore, the proposed ControlAR surprisingly empowers AR models with arbitrary-resolution image generation via conditional decoding and specific controls. Extensive experiments can demonstrate the controllability of the proposed ControlAR for the autoregressive control-to-image generation across diverse inputs, including edges, depths, and segmentation masks. Furthermore, both quantitative and qualitative results indicate that ControlAR surpasses previous state-of-the-art controllable diffusion models, e.g., ControlNet++. Code, models, and demo will soon be available at https://github.com/hustvl/ControlAR.

comment: Preprint. Work in progress

☆ Lie Algebra Canonicalization: Equivariant Neural Operators under arbitrary Lie Groups

Zakhar Shumaylov, Peter Zaika, James Rowbottom, Ferdia Sherry, Melanie Weber, Carola-Bibiane Schönlieb

The quest for robust and generalizable machine learning models has driven recent interest in exploiting symmetries through equivariant neural networks. In the context of PDE solvers, recent works have shown that Lie point symmetries can be a useful inductive bias for Physics-Informed Neural Networks (PINNs) through data and loss augmentation. Despite this, directly enforcing equivariance within the model architecture for these problems remains elusive. This is because many PDEs admit non-compact symmetry groups, oftentimes not studied beyond their infinitesimal generators, making them incompatible with most existing equivariant architectures. In this work, we propose Lie aLgebrA Canonicalization (LieLAC), a novel approach that exploits only the action of infinitesimal generators of the symmetry group, circumventing the need for knowledge of the full group structure. To achieve this, we address existing theoretical issues in the canonicalization literature, establishing connections with frame averaging in the case of continuous non-compact groups. Operating within the framework of canonicalization, LieLAC can easily be integrated with unconstrained pre-trained models, transforming inputs to a canonical form before feeding them into the existing model, effectively aligning the input for model inference according to allowed symmetries. LieLAC utilizes standard Lie group descent schemes, achieving equivariance in pre-trained models. Finally, we showcase LieLAC's efficacy on tasks of invariant image classification and Lie point symmetry equivariant neural PDE solvers using pre-trained models.

comment: 40 pages; preprint

☆ Unsupervised Point Cloud Completion through Unbalanced Optimal Transport

Taekyung Lee, Jaemoo Choi, Jaewoong Choi

Unpaired point cloud completion explores methods for learning a completion map from unpaired incomplete and complete point cloud data. In this paper, we propose a novel approach for unpaired point cloud completion using the unbalanced optimal transport map, called Unbalanced Optimal Transport Map for Unpaired Point Cloud Completion (UOT-UPC). We demonstrate that the unpaired point cloud completion can be naturally interpreted as the Optimal Transport (OT) problem and introduce the Unbalanced Optimal Transport (UOT) approach to address the class imbalance problem, which is prevalent in unpaired point cloud completion datasets. Moreover, we analyze the appropriate cost function for unpaired completion tasks. This analysis shows that the InfoCD cost function is particularly well-suited for this task. Our model is the first attempt to leverage UOT for unpaired point cloud completion, achieving competitive or superior results on both single-category and multi-category datasets. In particular, our model is especially effective in scenarios with class imbalance, where the proportions of categories are different between the incomplete and complete point cloud datasets.

comment: 20 pages, 10 figures

☆ Measuring and Improving Persuasiveness of Generative Models

Somesh Singh, Yaman K Singla, Harini SI, Balaji Krishnamurthy

☆ Learning 3D Perception from Others' Predictions

Jinsu Yoo, Zhenyang Feng, Tai-Yu Pan, Yihong Sun, Cheng Perng Phoo, Xiangyu Chen, Mark Campbell, Kilian Q. Weinberger, Bharath Hariharan, Wei-Lun Chao

comment: Under review

☆ Why Sample Space Matters: Keyframe Sampling Optimization for LiDAR-based Place Recognition

Nikolaos Stathoulopoulos, Vidya Sumathy, Christoforos Kanellakis, George Nikolakopoulos

Recent advances in robotics are pushing real-world autonomy, enabling robots to perform long-term and large-scale missions. A crucial component for successful missions is the incorporation of loop closures through place recognition, which effectively mitigates accumulated pose estimation drift. Despite computational advancements, optimizing performance for real-time deployment remains challenging, especially in resource-constrained mobile robots and multi-robot systems since, conventional keyframe sampling practices in place recognition often result in retaining redundant information or overlooking relevant data, as they rely on fixed sampling intervals or work directly in the 3D space instead of the feature space. To address these concerns, we introduce the concept of sample space in place recognition and demonstrate how different sampling techniques affect the query process and overall performance. We then present a novel keyframe sampling approach for LiDAR-based place recognition, which focuses on redundancy minimization and information preservation in the hyper-dimensional descriptor space. This approach is applicable to both learning-based and handcrafted descriptors, and through the experimental validation across multiple datasets and descriptor frameworks, we demonstrate the effectiveness of our proposed method, showing it can jointly minimize redundancy and preserve essential information in real-time. The proposed approach maintains robust performance across various datasets without requiring parameter tuning, contributing to more efficient and reliable place recognition for a wide range of robotic applications.

comment: 20 pages, 15 figures. Submitted

☆ Diffusion-based Extreme Image Compression with Compressed Feature Initialization

Zhiyuan Li, Yanhui Zhou, Hao Wei, Chenyang Ge, Ajmal Mian

Diffusion-based extreme image compression methods have achieved impressive performance at extremely low bitrates. However, constrained by the iterative denoising process that starts from pure noise, these methods are limited in both fidelity and efficiency. To address these two issues, we present Relay Residual Diffusion Extreme Image Compression (RDEIC), which leverages compressed feature initialization and residual diffusion. Specifically, we first use the compressed latent features of the image with added noise, instead of pure noise, as the starting point to eliminate the unnecessary initial stages of the denoising process. Second, we design a novel relay residual diffusion that reconstructs the raw image by iteratively removing the added noise and the residual between the compressed and target latent features. Notably, our relay residual diffusion network seamlessly integrates pre-trained stable diffusion to leverage its robust generative capability for high-quality reconstruction. Third, we propose a fixed-step fine-tuning strategy to eliminate the discrepancy between the training and inference phases, further improving the reconstruction quality. Extensive experiments demonstrate that the proposed RDEIC achieves state-of-the-art visual quality and outperforms existing diffusion-based extreme image compression methods in both fidelity and efficiency. The source code will be provided in https://github.com/huai-chang/RDEIC.

☆ Spatial-Temporal Multi-Cuts for Online Multiple-Camera Vehicle Tracking

Fabian Herzog, Johannes Gilg, Philipp Wolters, Torben Teepe, Gerhard Rigoll

Accurate online multiple-camera vehicle tracking is essential for intelligent transportation systems, autonomous driving, and smart city applications. Like single-camera multiple-object tracking, it is commonly formulated as a graph problem of tracking-by-detection. Within this framework, existing online methods usually consist of two-stage procedures that cluster temporally first, then spatially, or vice versa. This is computationally expensive and prone to error accumulation. We introduce a graph representation that allows spatial-temporal clustering in a single, combined step: New detections are spatially and temporally connected with existing clusters. By keeping sparse appearance and positional cues of all detections in a cluster, our method can compare clusters based on the strongest available evidence. The final tracks are obtained online using a simple multicut assignment procedure. Our method does not require any training on the target scene, pre-extraction of single-camera tracks, or additional annotations. Notably, we outperform the online state-of-the-art on the CityFlow dataset in terms of IDF1 by more than 14%, and on the Synthehicle dataset by more than 25%, respectively. The code is publicly available.

☆ Plots Unlock Time-Series Understanding in Multimodal Models

Mayank Daswani, Mathias M. J. Bellaiche, Marc Wilson, Desislav Ivanov, Mikhail Papkov, Eva Schnider, Jing Tang, Kay Lamerigts, Gabriela Botea, Michael A. Sanchez, Yojan Patel, Shruthi Prabhakara, Shravya Shetty, Umesh Telang

While multimodal foundation models can now natively work with data beyond text, they remain underutilized in analyzing the considerable amounts of multi-dimensional time-series data in fields like healthcare, finance, and social sciences, representing a missed opportunity for richer, data-driven insights. This paper proposes a simple but effective method that leverages the existing vision encoders of these models to "see" time-series data via plots, avoiding the need for additional, potentially costly, model training. Our empirical evaluations show that this approach outperforms providing the raw time-series data as text, with the additional benefit that visual time-series representations demonstrate up to a 90% reduction in model API costs. We validate our hypothesis through synthetic data tasks of increasing complexity, progressing from simple functional form identification on clean data, to extracting trends from noisy scatter plots. To demonstrate generalizability from synthetic tasks with clear reasoning steps to more complex, real-world scenarios, we apply our approach to consumer health tasks - specifically fall detection, activity recognition, and readiness assessment - which involve heterogeneous, noisy data and multi-step reasoning. The overall success in plot performance over text performance (up to an 120% performance increase on zero-shot synthetic tasks, and up to 150% performance increase on real-world tasks), across both GPT and Gemini model families, highlights our approach's potential for making the best use of the native capabilities of foundation models.

comment: 49 pages

☆ Metrics Revolutions: Groundbreaking Insights into the Implementation of Metrics for Biomedical Image Segmentation

Gašper Podobnik, Tomaž Vrtovec

The evaluation of segmentation performance is a common task in biomedical image analysis, with its importance emphasized in the recently released metrics selection guidelines and computing frameworks. To quantitatively evaluate the alignment of two segmentations, researchers commonly resort to counting metrics, such as the Dice similarity coefficient, or distance-based metrics, such as the Hausdorff distance, which are usually computed by publicly available open-source tools with an inherent assumption that these tools provide consistent results. In this study we questioned this assumption, and performed a systematic implementation analysis along with quantitative experiments on real-world clinical data to compare 11 open-source tools for distance-based metrics computation against our highly accurate mesh-based reference implementation. The results revealed that statistically significant differences among all open-source tools are both surprising and concerning, since they question the validity of existing studies. Besides identifying the main sources of variation, we also provide recommendations for distance-based metrics computation.

☆ GI-GS: Global Illumination Decomposition on Gaussian Splatting for Inverse Rendering

Hongze Chen, Zehong Lin, Jun Zhang

We present GI-GS, a novel inverse rendering framework that leverages 3D Gaussian Splatting (3DGS) and deferred shading to achieve photo-realistic novel view synthesis and relighting. In inverse rendering, accurately modeling the shading processes of objects is essential for achieving high-fidelity results. Therefore, it is critical to incorporate global illumination to account for indirect lighting that reaches an object after multiple bounces across the scene. Previous 3DGS-based methods have attempted to model indirect lighting by characterizing indirect illumination as learnable lighting volumes or additional attributes of each Gaussian, while using baked occlusion to represent shadow effects. These methods, however, fail to accurately model the complex physical interactions between light and objects, making it impossible to construct realistic indirect illumination during relighting. To address this limitation, we propose to calculate indirect lighting using efficient path tracing with deferred shading. In our framework, we first render a G-buffer to capture the detailed geometry and material properties of the scene. Then, we perform physically-based rendering (PBR) only for direct lighting. With the G-buffer and previous rendering results, the indirect lighting can be calculated through a lightweight path tracing. Our method effectively models indirect lighting under any given lighting conditions, thereby achieving better novel view synthesis and relighting. Quantitative and qualitative results show that our GI-GS outperforms existing baselines in both rendering quality and efficiency.

☆ NL-Eye: Abductive NLI for Images

Mor Ventura, Michael Toker, Nitay Calderon, Zorik Gekhman, Yonatan Bitton, Roi Reichart

☆ High-Efficiency Neural Video Compression via Hierarchical Predictive Learning

Ming Lu, Zhihao Duan, Wuyang Cong, Dandan Ding, Fengqing Zhu, Zhan Ma

The enhanced Deep Hierarchical Video Compression-DHVC 2.0-has been introduced. This single-model neural video codec operates across a broad range of bitrates, delivering not only superior compression performance to representative methods but also impressive complexity efficiency, enabling real-time processing with a significantly smaller memory footprint on standard GPUs. These remarkable advancements stem from the use of hierarchical predictive coding. Each video frame is uniformly transformed into multiscale representations through hierarchical variational autoencoders. For a specific scale's feature representation of a frame, its corresponding latent residual variables are generated by referencing lower-scale spatial features from the same frame and then conditionally entropy-encoded using a probabilistic model whose parameters are predicted using same-scale temporal reference from previous frames and lower-scale spatial reference of the current frame. This feature-space processing operates from the lowest to the highest scale of each frame, completely eliminating the need for the complexity-intensive motion estimation and compensation techniques that have been standard in video codecs for decades. The hierarchical approach facilitates parallel processing, accelerating both encoding and decoding, and supports transmission-friendly progressive decoding, making it particularly advantageous for networked video applications in the presence of packet loss. Source codes will be made available.

☆ IC3M: In-Car Multimodal Multi-object Monitoring for Abnormal Status of Both Driver and Passengers

Zihan Fang, Zheng Lin, Senkang Hu, Hangcheng Cao, Yiqin Deng, Xianhao Chen, Yuguang Fang

Recently, in-car monitoring has emerged as a promising technology for detecting early-stage abnormal status of the driver and providing timely alerts to prevent traffic accidents. Although training models with multimodal data enhances the reliability of abnormal status detection, the scarcity of labeled data and the imbalance of class distribution impede the extraction of critical abnormal state features, significantly deteriorating training performance. Furthermore, missing modalities due to environment and hardware limitations further exacerbate the challenge of abnormal status identification. More importantly, monitoring abnormal health conditions of passengers, particularly in elderly care, is of paramount importance but remains underexplored. To address these challenges, we introduce our IC3M, an efficient camera-rotation-based multimodal framework for monitoring both driver and passengers in a car. Our IC3M comprises two key modules: an adaptive threshold pseudo-labeling strategy and a missing modality reconstruction. The former customizes pseudo-labeling thresholds for different classes based on the class distribution, generating class-balanced pseudo labels to guide model training effectively, while the latter leverages crossmodality relationships learned from limited labels to accurately recover missing modalities by distribution transferring from available modalities. Extensive experimental results demonstrate that IC3M outperforms state-of-the-art benchmarks in accuracy, precision, and recall while exhibiting superior robustness under limited labeled data and severe missing modality.

comment: 16 pages, 17 figures

☆ An Improved Variational Method for Image Denoising

Jing-En Huang, Jia-Wei Liao, Ku-Te Lin, Yu-Ju Tsai, Mei-Heng Yueh

The total variation (TV) method is an image denoising technique that aims to reduce noise by minimizing the total variation of the image, which measures the variation in pixel intensities. The TV method has been widely applied in image processing and computer vision for its ability to preserve edges and enhance image quality. In this paper, we propose an improved TV model for image denoising and the associated numerical algorithm to carry out the procedure, which is particularly effective in removing several types of noises and their combinations. Our improved model admits a unique solution and the associated numerical algorithm guarantees the convergence. Numerical experiments are demonstrated to show improved effectiveness and denoising quality compared to other TV models. Such encouraging results further enhance the utility of the TV method in image processing.

☆ Combining Pre- and Post-Demosaicking Noise Removal for RAW Video

Marco Sánchez-Beeckman, Antoni Buades, Nicola Brandonisio, Bilel Kanoun

Denoising is one of the fundamental steps of the processing pipeline that converts data captured by a camera sensor into a display-ready image or video. It is generally performed early in the pipeline, usually before demosaicking, although studies swapping their order or even conducting them jointly have been proposed. With the advent of deep learning, the quality of denoising algorithms has steadily increased. Even so, modern neural networks still have a hard time adapting to new noise levels and scenes, which is indispensable for real-world applications. With those in mind, we propose a self-similarity-based denoising scheme that weights both a pre- and a post-demosaicking denoiser for Bayer-patterned CFA video data. We show that a balance between the two leads to better image quality, and we empirically find that higher noise levels benefit from a higher influence pre-demosaicking. We also integrate temporal trajectory prefiltering steps before each denoiser, which further improve texture reconstruction. The proposed method only requires an estimation of the noise model at the sensor, accurately adapts to any noise level, and is competitive with the state of the art, making it suitable for real-world videography.

comment: 16 pages, 9 figures

☆ SuperGS: Super-Resolution 3D Gaussian Splatting via Latent Feature Field and Gradient-guided Splitting

Shiyun Xie, Zhiru Wang, Yinghao Zhu, Chengwei Pan

Recently, 3D Gaussian Splatting (3DGS) has exceled in novel view synthesis with its real-time rendering capabilities and superior quality. However, it faces challenges for high-resolution novel view synthesis (HRNVS) due to the coarse nature of primitives derived from low-resolution input views. To address this issue, we propose Super-Resolution 3DGS (SuperGS), which is an expansion of 3DGS designed with a two-stage coarse-to-fine training framework, utilizing pretrained low-resolution scene representation as an initialization for super-resolution optimization. Moreover, we introduce Multi-resolution Feature Gaussian Splatting (MFGS) to incorporates a latent feature field for flexible feature sampling and Gradient-guided Selective Splitting (GSS) for effective Gaussian upsampling. By integrating these strategies within the coarse-to-fine framework ensure both high fidelity and memory efficiency. Extensive experiments demonstrate that SuperGS surpasses state-of-the-art HRNVS methods on challenging real-world datasets using only low-resolution inputs.

☆ NestedMorph: Enhancing Deformable Medical Image Registration with Nested Attention Mechanisms WACV

Gurucharan Marthi Krishna Kumar, Janine Mendola, Amir Shmuel

Deformable image registration is crucial for aligning medical images in a non-linear fashion across different modalities, allowing for precise spatial correspondence between varying anatomical structures. This paper presents NestedMorph, a novel network utilizing a Nested Attention Fusion approach to improve intra-subject deformable registration between T1-weighted (T1w) MRI and diffusion MRI (dMRI) data. NestedMorph integrates high-resolution spatial details from an encoder with semantic information from a decoder using a multi-scale framework, enhancing both local and global feature extraction. Our model notably outperforms existing methods, including CNN-based approaches like VoxelMorph, MIDIR, and CycleMorph, as well as Transformer-based models such as TransMorph and ViT-V-Net, and traditional techniques like NiftyReg and SyN. Evaluations on the HCP dataset demonstrate that NestedMorph achieves superior performance across key metrics, including SSIM, HD95, and SDlogJ, with the highest SSIM of 0.89, and the lowest HD95 of 2.5 and SDlogJ of 0.22. These results highlight NestedMorph's ability to capture both local and global image features effectively, leading to superior registration performance. The promising outcomes of this study underscore NestedMorph's potential to significantly advance deformable medical image registration, providing a robust framework for future research and clinical applications. The source code and our implementation are available at: https://bit.ly/3zdVqcg

comment: Submitted to IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025

☆ MedVisionLlama: Leveraging Pre-Trained Large Language Model Layers to Enhance Medical Image Segmentation WACV

Gurucharan Marthi Krishna Kumar, Aman Chadha, Janine Mendola, Amir Shmuel

comment: Submitted to IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025

☆ Pseudo-Stereo Inputs: A Solution to the Occlusion Challenge in Self-Supervised Stereo Matching

Ruizhi Yang, Xingqiang Li, Jiajun Bai, Jinsong Du

Self-supervised stereo matching holds great promise for application and research due to its independence from expensive labeled data. However, direct self-supervised stereo matching paradigms based on photometric loss functions have consistently struggled with performance issues due to the occlusion challenge. The crux of the occlusion challenge lies in the fact that the positions of occluded pixels consistently align with the epipolar search direction defined by the input stereo images, leading to persistent information loss and erroneous feedback at fixed locations during self-supervised training. In this work, we propose a simple yet highly effective pseudo-stereo inputs strategy to address the core occlusion challenge. This strategy decouples the input and feedback images, compelling the network to probabilistically sample information from both sides of the occluding objects. As a result, the persistent lack of information in the aforementioned fixed occlusion areas is mitigated. Building upon this, we further address feedback conflicts and overfitting issues arising from the strategy. By integrating these components, our method achieves stable and significant performance improvements compared to existing methods. Quantitative experiments are conducted to evaluate the performance. Qualitative experiments further demonstrate accurate disparity inference even at occluded regions. These results demonstrate a significant advancement over previous methods in the field of direct self-supervised stereo matching based on photometric loss. The proposed pseudo-stereo inputs strategy, due to its simplicity and effectiveness, has the potential to serve as a new paradigm for direct self-supervised stereo matching. Code is available at https://github.com/qrzyang/Pseudo-Stereo.

comment: Submitted to IEEE Transactions on Image Processing (TIP)

☆ A Foundation Model for the Solar Dynamics Observatory

James Walsh, Daniel G. Gass, Raul Ramos Pollan, Paul J. Wright, Richard Galvez, Noah Kasmanoff, Jason Naradowsky, Anne Spalding, James Parr, Atılım Güneş Baydin

SDO-FM is a foundation model using data from NASA's Solar Dynamics Observatory (SDO) spacecraft; integrating three separate instruments to encapsulate the Sun's complex physical interactions into a multi-modal embedding space. This model can be used to streamline scientific investigations involving SDO by making the enormous datasets more computationally accessible for heliophysics research and enable investigations that require instrument fusion. We discuss four key components: an ingestion pipeline to create machine learning ready datasets, the model architecture and training approach, resultant embeddings and fine-tunable models, and finally downstream fine-tuned applications. A key component of this effort has been to include subject matter specialists at each stage of development; reviewing the scientific value and providing guidance for model architecture, dataset, and training paradigm decisions. This paper marks release of our pretrained models and embedding datasets, available to the community on Hugging Face and sdofm.org.

☆ HiFiSeg: High-Frequency Information Enhanced Polyp Segmentation with Global-Local Vision Transformer

Jingjing Ren, Xiaoyong Zhang, Lina Zhang

Numerous studies have demonstrated the strong performance of Vision Transformer (ViT)-based methods across various computer vision tasks. However, ViT models often struggle to effectively capture high-frequency components in images, which are crucial for detecting small targets and preserving edge details, especially in complex scenarios. This limitation is particularly challenging in colon polyp segmentation, where polyps exhibit significant variability in structure, texture, and shape. High-frequency information, such as boundary details, is essential for achieving precise semantic segmentation in this context. To address these challenges, we propose HiFiSeg, a novel network for colon polyp segmentation that enhances high-frequency information processing through a global-local vision transformer framework. HiFiSeg leverages the pyramid vision transformer (PVT) as its encoder and introduces two key modules: the global-local interaction module (GLIM) and the selective aggregation module (SAM). GLIM employs a parallel structure to fuse global and local information at multiple scales, effectively capturing fine-grained features. SAM selectively integrates boundary details from low-level features with semantic information from high-level features, significantly improving the model's ability to accurately detect and segment polyps. Extensive experiments on five widely recognized benchmark datasets demonstrate the effectiveness of HiFiSeg for polyp segmentation. Notably, the mDice scores on the challenging CVC-ColonDB and ETIS datasets reached 0.826 and 0.822, respectively, underscoring the superior performance of HiFiSeg in handling the specific complexities of this task.

☆ Learning from Offline Foundation Features with Tensor Augmentations NeurIPS 2024

Emir Konuk, Christos Matsoukas, Moein Sorkhei, Phitchapha Lertsiravaramet, Kevin Smith

We introduce Learning from Offline Foundation Features with Tensor Augmentations (LOFF-TA), an efficient training scheme designed to harness the capabilities of foundation models in limited resource settings where their direct development is not feasible. LOFF-TA involves training a compact classifier on cached feature embeddings from a frozen foundation model, resulting in up to $37\times$ faster training and up to $26\times$ reduced GPU memory usage. Because the embeddings of augmented images would be too numerous to store, yet the augmentation process is essential for training, we propose to apply tensor augmentations to the cached embeddings of the original non-augmented images. LOFF-TA makes it possible to leverage the power of foundation models, regardless of their size, in settings with limited computational capacity. Moreover, LOFF-TA can be used to apply foundation models to high-resolution images without increasing compute. In certain scenarios, we find that training with LOFF-TA yields better results than directly fine-tuning the foundation model.

comment: Accepted to the 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

☆ Med-TTT: Vision Test-Time Training model for Medical Image Segmentation

Jiashu Xu

Medical image segmentation plays a crucial role in clinical diagnosis and treatment planning. Although models based on convolutional neural networks (CNNs) and Transformers have achieved remarkable success in medical image segmentation tasks, they still face challenges such as high computational complexity and the loss of local features when capturing long-range dependencies. To address these limitations, we propose Med-TTT, a visual backbone network integrated with Test-Time Training (TTT) layers, which incorporates dynamic adjustment capabilities. Med-TTT introduces the Vision-TTT layer, which enables effective modeling of long-range dependencies with linear computational complexity and adaptive parameter adjustment during inference. Furthermore, we designed a multi-resolution fusion mechanism to combine image features at different scales, facilitating the identification of subtle lesion characteristics in complex backgrounds. At the same time, we adopt a frequency domain feature enhancement strategy based on high pass filtering, which can better capture texture and fine-grained details in images. Experimental results demonstrate that Med-TTT significantly outperforms existing methods on multiple medical image datasets, exhibiting strong segmentation capabilities, particularly in complex image backgrounds. The model achieves leading performance in terms of accuracy, sensitivity, and Dice coefficient, providing an efficient and robust solution for the field of medical image segmentation.The code is available at https://github.com/Jiashu-Xu/Med-TTT .

☆ Dog-IQA: Standard-guided Zero-shot MLLM for Mix-grained Image Quality Assessment

Kai Liu, Ziqing Zhang, Wenbo Li, Renjing Pei, Fenglong Song, Xiaohong Liu, Linghe Kong, Yulun Zhang

Image quality assessment (IQA) serves as the golden standard for all models' performance in nearly all computer vision fields. However, it still suffers from poor out-of-distribution generalization ability and expensive training costs. To address these problems, we propose Dog-IQA, a standard-guided zero-shot mix-grained IQA method, which is training-free and utilizes the exceptional prior knowledge of multimodal large language models (MLLMs). To obtain accurate IQA scores, namely scores consistent with humans, we design an MLLM-based inference pipeline that imitates human experts. In detail, Dog-IQA applies two techniques. First, Dog-IQA objectively scores with specific standards that utilize MLLM's behavior pattern and minimize the influence of subjective factors. Second, Dog-IQA comprehensively takes local semantic objects and the whole image as input and aggregates their scores, leveraging local and global information. Our proposed Dog-IQA achieves state-of-the-art (SOTA) performance compared with training-free methods, and competitive performance compared with training-based methods in cross-dataset scenarios. Our code and models will be available at https://github.com/Kai-Liu001/Dog-IQA.

comment: 10 pages, 5 figures. The code and models will be available at https://github.com/Kai-Liu001/Dog-IQA

☆ DTVLT: A Multi-modal Diverse Text Benchmark for Visual Language Tracking Based on LLM

Xuchen Li, Shiyu Hu, Xiaokun Feng, Dailing Zhang, Meiqi Wu, Jing Zhang, Kaiqi Huang

comment: Preprint, Under Review

☆ Event-Customized Image Generation

Zhen Wang, Yilei Jiang, Dong Zheng, Jun Xiao, Long Chen

Customized Image Generation, generating customized images with user-specified concepts, has raised significant attention due to its creativity and novelty. With impressive progress achieved in subject customization, some pioneer works further explored the customization of action and interaction beyond entity (i.e., human, animal, and object) appearance. However, these approaches only focus on basic actions and interactions between two entities, and their effects are limited by insufficient ''exactly same'' reference images. To extend customized image generation to more complex scenes for general real-world applications, we propose a new task: event-customized image generation. Given a single reference image, we define the ''event'' as all specific actions, poses, relations, or interactions between different entities in the scene. This task aims at accurately capturing the complex event and generating customized images with various target entities. To solve this task, we proposed a novel training-free event customization method: FreeEvent. Specifically, FreeEvent introduces two extra paths alongside the general diffusion denoising process: 1) Entity switching path: it applies cross-attention guidance and regulation for target entity generation. 2) Event transferring path: it injects the spatial feature and self-attention maps from the reference image to the target image for event generation. To further facilitate this new task, we collected two evaluation benchmarks: SWiG-Event and Real-Event. Extensive experiments and ablations have demonstrated the effectiveness of FreeEvent.

☆ Towards a Theoretical Understanding of Memorization in Diffusion Models

Yunhao Chen, Xingjun Ma, Difan Zou, Yu-Gang Jiang

As diffusion probabilistic models (DPMs) are being employed as mainstream models for Generative Artificial Intelligence (GenAI), the study of their memorization of training data has attracted growing attention. Existing works in this direction aim to establish an understanding of whether or to what extent DPMs learn via memorization. Such an understanding is crucial for identifying potential risks of data leakage and copyright infringement in diffusion models and, more importantly, for trustworthy application of GenAI. Existing works revealed that conditional DPMs are more prone to training data memorization than unconditional DPMs, and the motivated data extraction methods are mostly for conditional DPMs. However, these understandings are primarily empirical, and extracting training data from unconditional models has been found to be extremely challenging. In this work, we provide a theoretical understanding of memorization in both conditional and unconditional DPMs under the assumption of model convergence. Our theoretical analysis indicates that extracting data from unconditional models can also be effective by constructing a proper surrogate condition. Based on this result, we propose a novel data extraction method named \textbf{Surrogate condItional Data Extraction (SIDE)} that leverages a time-dependent classifier trained on the generated data as a surrogate condition to extract training data from unconditional DPMs. Empirical results demonstrate that our SIDE can extract training data in challenging scenarios where previous methods fail, and it is, on average, over 50\% more effective across different scales of the CelebA dataset.

comment: arXiv admin note: text overlap with arXiv:2406.12752

☆ Recurrent Few-Shot model for Document Verification

Maxime Talarmain, Carlos Boned, Sanket Biswas, Oriol Ramos

General-purpose ID, or travel, document image- and video-based verification systems have yet to achieve good enough performance to be considered a solved problem. There are several factors that negatively impact their performance, including low-resolution images and videos and a lack of sufficient data to train the models. This task is particularly challenging when dealing with unseen class of ID, or travel, documents. In this paper we address this task by proposing a recurrent-based model able to detect forged documents in a few-shot scenario. The recurrent architecture makes the model robust to document resolution variability. Moreover, the few-shot approach allow the model to perform well even for unseen class of documents. Preliminary results on the SIDTD and Findit datasets show good performance of this model for this task.

☆ Clinnova Federated Learning Proof of Concept: Key Takeaways from a Cross-border Collaboration

Julia Alekseenko, Bram Stieltjes, Michael Bach, Melanie Boerries, Oliver Opitz, Alexandros Karargyris, Nicolas Padoy

Clinnova, a collaborative initiative involving France, Germany, Switzerland, and Luxembourg, is dedicated to unlocking the power of precision medicine through data federation, standardization, and interoperability. This European Greater Region initiative seeks to create an interoperable European standard using artificial intelligence (AI) and data science to enhance healthcare outcomes and efficiency. Key components include multidisciplinary research centers, a federated biobanking strategy, a digital health innovation platform, and a federated AI strategy. It targets inflammatory bowel disease, rheumatoid diseases, and multiple sclerosis (MS), emphasizing data quality to develop AI algorithms for personalized treatment and translational research. The IHU Strasbourg (Institute of Minimal-invasive Surgery) has the lead in this initiative to develop the federated learning (FL) proof of concept (POC) that will serve as a foundation for advancing AI in healthcare. At its core, Clinnova-MS aims to enhance MS patient care by using FL to develop more accurate models that detect disease progression, guide interventions, and validate digital biomarkers across multiple sites. This technical report presents insights and key takeaways from the first cross-border federated POC on MS segmentation of MRI images within the Clinnova framework. While our work marks a significant milestone in advancing MS segmentation through cross-border collaboration, it also underscores the importance of addressing technical, logistical, and ethical considerations to realize the full potential of FL in healthcare settings.

☆ Predictive Attractor Models NeurIPS 2024

Ramy Mounir, Sudeep Sarkar

Sequential memory, the ability to form and accurately recall a sequence of events or stimuli in the correct order, is a fundamental prerequisite for biological and artificial intelligence as it underpins numerous cognitive functions (e.g., language comprehension, planning, episodic memory formation, etc.) However, existing methods of sequential memory suffer from catastrophic forgetting, limited capacity, slow iterative learning procedures, low-order Markov memory, and, most importantly, the inability to represent and generate multiple valid future possibilities stemming from the same context. Inspired by biologically plausible neuroscience theories of cognition, we propose \textit{Predictive Attractor Models (PAM)}, a novel sequence memory architecture with desirable generative properties. PAM is a streaming model that learns a sequence in an online, continuous manner by observing each input \textit{only once}. Additionally, we find that PAM avoids catastrophic forgetting by uniquely representing past context through lateral inhibition in cortical minicolumns, which prevents new memories from overwriting previously learned knowledge. PAM generates future predictions by sampling from a union set of predicted possibilities; this generative ability is realized through an attractor model trained alongside the predictor. We show that PAM is trained with local computations through Hebbian plasticity rules in a biologically plausible framework. Other desirable traits (e.g., noise tolerance, CPU-based learning, capacity scaling) are discussed throughout the paper. Our findings suggest that PAM represents a significant step forward in the pursuit of biologically plausible and computationally efficient sequential memory models, with broad implications for cognitive science and artificial intelligence research.

comment: Accepted to NeurIPS 2024

☆ PnP-Flow: Plug-and-Play Image Restoration with Flow Matching

Ségolène Martin, Anne Gagneux, Paul Hagemann, Gabriele Steidl

In this paper, we introduce Plug-and-Play (PnP) Flow Matching, an algorithm for solving imaging inverse problems. PnP methods leverage the strength of pre-trained denoisers, often deep neural networks, by integrating them in optimization schemes. While they achieve state-of-the-art performance on various inverse problems in imaging, PnP approaches face inherent limitations on more generative tasks like inpainting. On the other hand, generative models such as Flow Matching pushed the boundary in image sampling yet lack a clear method for efficient use in image restoration. We propose to combine the PnP framework with Flow Matching (FM) by defining a time-dependent denoiser using a pre-trained FM model. Our algorithm alternates between gradient descent steps on the data-fidelity term, reprojections onto the learned FM path, and denoising. Notably, our method is computationally efficient and memory-friendly, as it avoids backpropagation through ODEs and trace computations. We evaluate its performance on denoising, super-resolution, deblurring, and inpainting tasks, demonstrating superior results compared to existing PnP algorithms and Flow Matching based state-of-the-art methods.

☆ LoGDesc: Local geometric features aggregation for robust point cloud registration

Karim Slimani, Brahim Tamadazte, Catherine Achard

This paper introduces a new hybrid descriptor for 3D point matching and point cloud registration, combining local geometrical properties and learning-based feature propagation for each point's neighborhood structure description. The proposed architecture first extracts prior geometrical information by computing each point's planarity, anisotropy, and omnivariance using a Principal Components Analysis (PCA). This prior information is completed by a descriptor based on the normal vectors estimated thanks to constructing a neighborhood based on triangles. The final geometrical descriptor is propagated between the points using local graph convolutions and attention mechanisms. The new feature extractor is evaluated on ModelNet40, Bunny Stanford dataset, KITTI and MVP (Multi-View Partial)-RG for point cloud registration and shows interesting results, particularly on noisy and low overlapping point clouds.

☆ Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models

Seyedmorteza Sadat, Otmar Hilliges, Romann M. Weber

Classifier-free guidance (CFG) is crucial for improving both generation quality and alignment between the input condition and final output in diffusion models. While a high guidance scale is generally required to enhance these aspects, it also causes oversaturation and unrealistic artifacts. In this paper, we revisit the CFG update rule and introduce modifications to address this issue. We first decompose the update term in CFG into parallel and orthogonal components with respect to the conditional model prediction and observe that the parallel component primarily causes oversaturation, while the orthogonal component enhances image quality. Accordingly, we propose down-weighting the parallel component to achieve high-quality generations without oversaturation. Additionally, we draw a connection between CFG and gradient ascent and introduce a new rescaling and momentum method for the CFG update rule based on this insight. Our approach, termed adaptive projected guidance (APG), retains the quality-boosting advantages of CFG while enabling the use of higher guidance scales without oversaturation. APG is easy to implement and introduces practically no additional computational overhead to the sampling process. Through extensive experiments, we demonstrate that APG is compatible with various conditional diffusion models and samplers, leading to improved FID, recall, and saturation scores while maintaining precision comparable to CFG, making our method a superior plug-and-play alternative to standard classifier-free guidance.

☆ SynCo: Synthetic Hard Negatives in Contrastive Learning for Better Unsupervised Visual Representations

Nikolaos Giakoumoglou, Tania Stathaki

Contrastive learning has become a dominant approach in self-supervised visual representation learning, with hard negatives-samples that closely resemble the anchor-being key to enhancing the discriminative power of learned representations. However, efficiently leveraging hard negatives remains a challenge due to the difficulty in identifying and incorporating them without significantly increasing computational costs. To address this, we introduce SynCo (Synthetic Negatives in Contrastive learning), a novel contrastive learning approach that improves model performance by generating synthetic hard negatives. Built on the MoCo framework, SynCo introduces six novel strategies for creating diverse synthetic hard negatives that can be generated on-the-fly with minimal computational overhead. SynCo achieves faster training and better representation learning, achieving a top-1 accuracy of 68.1% in ImageNet linear evaluation after only 200 epochs on pretraining, surpassing MoCo's 67.5% with the same ResNet-50 encoder. Additionally, it transfers more effectively to detection tasks: on the PASCAL VOC, it outperforms both the supervised baseline and MoCo, achieving an AP of 82.5%; on the COCO dataset, it sets a new benchmark with 40.4% AP for bounding box detection and 35.4% AP for instance segmentation. Our synthetic hard negative generation procedure significantly enhances the quality of visual representations learned through self-supervised contrastive learning. Code is available at https://github.com/giakoumoglou/synco.

comment: 10 pages, 6 figures, 4 tables. arXiv admin note: text overlap with arXiv:2010.01028 by other authors

☆ Parameter Competition Balancing for Model Merging NeurIPS2024

Guodong Du, Junlin Lee, Jing Li, Runhua Jiang, Yifei Guo, Shuyang Yu, Hanting Liu, Sim Kuan Goh, Ho-Kin Tang, Daojing He, Min Zhang

comment: Accepted by NeurIPS2024

☆ MetaMetrics: Calibrating Metrics For Generation Tasks Using Human Preferences

Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, Derry Tanti Wijaya

comment: Preprint

☆ Unleashing the Potential of the Diffusion Model in Few-shot Semantic Segmentation NeurIPS

Muzhi Zhu, Yang Liu, Zekai Luo, Chenchen Jing, Hao Chen, Guangkai Xu, Xinlong Wang, Chunhua Shen

comment: Accepted to Proc. Annual Conference on Neural Information Processing Systems (NeurIPS) 2024

☆ A Comprehensive Survey of Mamba Architectures for Medical Image Analysis: Classification, Segmentation, Restoration and Beyond

Shubhi Bansal, Sreeharish A, Madhava Prasath J, Manikandan S, Sreekanth Madisetty, Mohammad Zia Ur Rehman, Chandravardhan Singh Raghaw, Gaurav Duggal, Nagendra Kumar

Mamba, a special case of the State Space Model, is gaining popularity as an alternative to template-based deep learning approaches in medical image analysis. While transformers are powerful architectures, they have drawbacks, including quadratic computational complexity and an inability to address long-range dependencies efficiently. This limitation affects the analysis of large and complex datasets in medical imaging, where there are many spatial and temporal relationships. In contrast, Mamba offers benefits that make it well-suited for medical image analysis. It has linear time complexity, which is a significant improvement over transformers. Mamba processes longer sequences without attention mechanisms, enabling faster inference and requiring less memory. Mamba also demonstrates strong performance in merging multimodal data, improving diagnosis accuracy and patient outcomes. The organization of this paper allows readers to appreciate the capabilities of Mamba in medical imaging step by step. We begin by defining core concepts of SSMs and models, including S4, S5, and S6, followed by an exploration of Mamba architectures such as pure Mamba, U-Net variants, and hybrid models with convolutional neural networks, transformers, and Graph Neural Networks. We also cover Mamba optimizations, techniques and adaptations, scanning, datasets, applications, experimental results, and conclude with its challenges and future directions in medical imaging. This review aims to demonstrate the transformative potential of Mamba in overcoming existing barriers within medical imaging while paving the way for innovative advancements in the field. A comprehensive list of Mamba architectures applied in the medical field, reviewed in this work, is available at Github.

☆ ProtoSeg: A Prototype-Based Point Cloud Instance Segmentation Method

Remco Royen, Leon Denis, Adrian Munteanu

3D instance segmentation is crucial for obtaining an understanding of a point cloud scene. This paper presents a novel neural network architecture for performing instance segmentation on 3D point clouds. We propose to jointly learn coefficients and prototypes in parallel which can be combined to obtain the instance predictions. The coefficients are computed using an overcomplete set of sampled points with a novel multi-scale module, dubbed dilated point inception. As the set of obtained instance mask predictions is overcomplete, we employ a non-maximum suppression algorithm to retrieve the final predictions. This approach allows to omit the time-expensive clustering step and leads to a more stable inference time. The proposed method is not only 28% faster than the state-of-the-art, it also exhibits the lowest standard deviation. Our experiments have shown that the standard deviation of the inference time is only 1.0% of the total time while it ranges between 10.8 and 53.1% for the state-of-the-art methods. Lastly, our method outperforms the state-of-the-art both on S3DIS-blocks (4.9% in mRec on Fold-5) and PartNet (2.0% on average in mAP).

☆ Self-eXplainable AI for Medical Image Analysis: A Survey and New Outlooks

Junlin Hou, Sicen Liu, Yequan Bie, Hongmei Wang, Andong Tan, Luyang Luo, Hao Chen

The increasing demand for transparent and reliable models, particularly in high-stakes decision-making areas such as medical image analysis, has led to the emergence of eXplainable Artificial Intelligence (XAI). Post-hoc XAI techniques, which aim to explain black-box models after training, have been controversial in recent works concerning their fidelity to the models' predictions. In contrast, Self-eXplainable AI (S-XAI) offers a compelling alternative by incorporating explainability directly into the training process of deep learning models. This approach allows models to generate inherent explanations that are closely aligned with their internal decision-making processes. Such enhanced transparency significantly supports the trustworthiness, robustness, and accountability of AI systems in real-world medical applications. To facilitate the development of S-XAI methods for medical image analysis, this survey presents an comprehensive review across various image modalities and clinical applications. It covers more than 200 papers from three key perspectives: 1) input explainability through the integration of explainable feature engineering and knowledge graph, 2) model explainability via attention-based learning, concept-based learning, and prototype-based learning, and 3) output explainability by providing counterfactual explanation and textual explanation. Additionally, this paper outlines the desired characteristics of explainability and existing evaluation methods for assessing explanation quality. Finally, it discusses the major challenges and future research directions in developing S-XAI for medical image analysis.

☆ RESSCAL3D++: Joint Acquisition and Semantic Segmentation of 3D Point Clouds ICIP

Remco Royen, Kostas Pataridis, Ward van der Tempel, Adrian Munteanu

3D scene understanding is crucial for facilitating seamless interaction between digital devices and the physical world. Real-time capturing and processing of the 3D scene are essential for achieving this seamless integration. While existing approaches typically separate acquisition and processing for each frame, the advent of resolution-scalable 3D sensors offers an opportunity to overcome this paradigm and fully leverage the otherwise wasted acquisition time to initiate processing. In this study, we introduce VX-S3DIS, a novel point cloud dataset accurately simulating the behavior of a resolution-scalable 3D sensor. Additionally, we present RESSCAL3D++, an important improvement over our prior work, RESSCAL3D, by incorporating an update module and processing strategy. By applying our method to the new dataset, we practically demonstrate the potential of joint acquisition and semantic segmentation of 3D point clouds. Our resolution-scalable approach significantly reduces scalability costs from 2% to just 0.2% in mIoU while achieving impressive speed-ups of 15.6 to 63.9% compared to the non-scalable baseline. Furthermore, our scalable approach enables early predictions, with the first one occurring after only 7% of the total inference time of the baseline. The new VX-S3DIS dataset is available at https://github.com/remcoroyen/vx-s3dis.

comment: 2024 IEEE International Conference on Image Processing (ICIP). IEEE, 2024

☆ CTARR: A fast and robust method for identifying anatomical regions on CT images via atlas registration

Thomas Buddenkotte, Roland Opfer, Julia Krüger, Alessa Hering, Mireia Crispin-Ortuzar

Medical image analysis tasks often focus on regions or structures located in a particular location within the patient's body. Often large parts of the image may not be of interest for the image analysis task. When using deep-learning based approaches, this causes an unnecessary increases the computational burden during inference and raises the chance of errors. In this paper, we introduce CTARR, a novel generic method for CT Anatomical Region Recognition. The method serves as a pre-processing step for any deep learning-based CT image analysis pipeline by automatically identifying the pre-defined anatomical region that is relevant for the follow-up task and removing the rest. It can be used in (i) image segmentation to prevent false positives in anatomically implausible regions and speeding up the inference, (ii) image classification to produce image crops that are consistent in their anatomical context, and (iii) image registration by serving as a fast pre-registration step. Our proposed method is based on atlas registration and provides a fast and robust way to crop any anatomical region encoded as one or multiple bounding box(es) from any unlabeled CT scan of the brain, chest, abdomen and/or pelvis. We demonstrate the utility and robustness of the proposed method in the context of medical image segmentation by evaluating it on six datasets of public segmentation challenges. The foreground voxels in the regions of interest are preserved in the vast majority of cases and tasks (97.45-100%) while taking only fractions of a seconds to compute (0.1-0.21s) on a deep learning workstation and greatly reducing the segmentation runtime (2.0-12.7x). Our code is available at https://github.com/ThomasBudd/ctarr.

☆ Decoupling Layout from Glyph in Online Chinese Handwriting Generation

Ren-Min Si, Yan-Ming Zhang, Yi Chen

☆ The Comparison of Individual Cat Recognition Using Neural Networks

Mingxuan Li, Kai Zhou

Facial recognition using deep learning has been widely used in social life for applications such as authentication, smart door locks, and photo grouping, etc. More and more networks have been developed to facilitate computer vision tasks, such as ResNet, DenseNet, EfficientNet, ConvNeXt, and Siamese networks. However, few studies have systematically compared the advantages and disadvantages of such neural networks in identifying individuals from images, especially for pet animals like cats. In the present study, by systematically comparing the efficacy of different neural networks in cat recognition, we found traditional CNNs trained with transfer learning have better performance than models trained with the fine-tuning method or Siamese networks in individual cat recognition. In addition, ConvNeXt and DenseNet yield significant results which could be further optimized for individual cat recognition in pet stores and in the wild. These results provide a method to improve cat management in pet stores and monitoring of cats in the wild.

comment: 13 pages,7 figures

☆ A Novel Method for Accurate & Real-time Food Classification: The Synergistic Integration of EfficientNetB7, CBAM, Transfer Learning, and Data Augmentation

Shayan Rokhva, Babak Teimourpour

Integrating artificial intelligence into modern society is profoundly transformative, significantly enhancing productivity by streamlining various daily tasks. AI-driven recognition systems provide notable advantages in the food sector, including improved nutrient tracking, tackling food waste, and boosting food production and consumption efficiency. Accurate food classification is a crucial initial step in utilizing advanced AI models, as the effectiveness of this process directly influences the success of subsequent operations; therefore, achieving high accuracy at a reasonable speed is essential. Despite existing research efforts, a gap persists in improving performance while ensuring rapid processing times, prompting researchers to pursue cost-effective and precise models. This study addresses this gap by employing the state-of-the-art EfficientNetB7 architecture, enhanced through transfer learning, data augmentation, and the CBAM attention module. This methodology results in a robust model that surpasses previous studies in accuracy while maintaining rapid processing suitable for real-world applications. The Food11 dataset from Kaggle was utilized, comprising 16643 imbalanced images across 11 diverse classes with significant intra-category diversities and inter-category similarities. Furthermore, the proposed methodology, bolstered by various deep learning techniques, consistently achieves an impressive average accuracy of 96.40%. Notably, it can classify over 60 images within one second during inference on unseen data, demonstrating its ability to deliver high accuracy promptly. This underscores its potential for practical applications in accurate food classification and enhancing efficiency in subsequent processes.

comment: 20 pages, six figures, two tables

☆ Computer-aided Colorization State-of-the-science: A Survey

Yu Cao, Xin Duan, Xiangqiao Meng, P. Y. Mok, Ping Li, Tong-Yee Lee

This paper reviews published research in the field of computer-aided colorization technology. We argue that the colorization task originates from computer graphics, prospers by introducing computer vision, and tends to the fusion of vision and graphics, so we put forward our taxonomy and organize the whole paper chronologically. We extend the existing reconstruction-based colorization evaluation techniques, considering that aesthetic assessment of colored images should be introduced to ensure that colorization satisfies human visual-related requirements and emotions more closely. We perform the colorization aesthetic assessment on seven representative unconditional colorization models and discuss the difference between our assessment and the existing reconstruction-based metrics. Finally, this paper identifies unresolved issues and proposes fruitful areas for future research and development. Access to the project associated with this survey can be obtained at https://github.com/DanielCho-HK/Colorization.

☆ Structural-Entropy-Based Sample Selection for Efficient and Effective Learning ICLR 2025

Tianchi Xie, Jiangning Zhu, Guozu Ma, Minzhi Lin, Wei Chen, Weikai Yang, Shixia Liu

Sample selection improves the efficiency and effectiveness of machine learning models by providing informative and representative samples. Typically, samples can be modeled as a sample graph, where nodes are samples and edges represent their similarities. Most existing methods are based on local information, such as the training difficulty of samples, thereby overlooking global information, such as connectivity patterns. This oversight can result in suboptimal selection because global information is crucial for ensuring that the selected samples well represent the structural properties of the graph. To address this issue, we employ structural entropy to quantify global information and losslessly decompose it from the whole graph to individual nodes using the Shapley value. Based on the decomposition, we present $\textbf{S}$tructural-$\textbf{E}$ntropy-based sample $\textbf{S}$election ($\textbf{SES}$), a method that integrates both global and local information to select informative and representative samples. SES begins by constructing a $k$NN-graph among samples based on their similarities. It then measures sample importance by combining structural entropy (global metric) with training difficulty (local metric). Finally, SES applies importance-biased blue noise sampling to select a set of diverse and representative samples. Comprehensive experiments on three learning scenarios -- supervised learning, active learning, and continual learning -- clearly demonstrate the effectiveness of our method.

comment: Submitted to ICLR 2025

☆ Probabilistic road classification in historical maps using synthetic data and deep learning

Dominik J. Mühlematter, Sebastian Schweizer, Chenjing Jiao, Xue Xia, Magnus Heitzler, Lorenz Hurni

Historical maps are invaluable for analyzing long-term changes in transportation and spatial development, offering a rich source of data for evolutionary studies. However, digitizing and classifying road networks from these maps is often expensive and time-consuming, limiting their widespread use. Recent advancements in deep learning have made automatic road extraction from historical maps feasible, yet these methods typically require large amounts of labeled training data. To address this challenge, we introduce a novel framework that integrates deep learning with geoinformation, computer-based painting, and image processing methodologies. This framework enables the extraction and classification of roads from historical maps using only road geometries without needing road class labels for training. The process begins with training of a binary segmentation model to extract road geometries, followed by morphological operations, skeletonization, vectorization, and filtering algorithms. Synthetic training data is then generated by a painting function that artificially re-paints road segments using predefined symbology for road classes. Using this synthetic data, a deep ensemble is trained to generate pixel-wise probabilities for road classes to mitigate distribution shift. These predictions are then discretized along the extracted road geometries. Subsequently, further processing is employed to classify entire roads, enabling the identification of potential changes in road classes and resulting in a labeled road class dataset. Our method achieved completeness and correctness scores of over 94% and 92%, respectively, for road class 2, the most prevalent class in the two Siegfried Map sheets from Switzerland used for testing. This research offers a powerful tool for urban planning and transportation decision-making by efficiently extracting and classifying roads from historical maps.

☆ Spiking Neural Network as Adaptive Event Stream Slicer NeurIPS 2024

Jiahang Cao, Mingyuan Sun, Ziqing Wang, Hao Cheng, Qiang Zhang, Shibo Zhou, Renjing Xu

Event-based cameras are attracting significant interest as they provide rich edge information, high dynamic range, and high temporal resolution. Many state-of-the-art event-based algorithms rely on splitting the events into fixed groups, resulting in the omission of crucial temporal information, particularly when dealing with diverse motion scenarios (e.g., high/low speed). In this work, we propose SpikeSlicer, a novel-designed plug-and-play event processing method capable of splitting events stream adaptively. SpikeSlicer utilizes a lightweight (0.41M) and low-energy spiking neural network (SNN) to trigger event slicing. To guide the SNN to fire spikes at optimal time steps, we propose the Spiking Position-aware Loss (SPA-Loss) to modulate the neuron's state. Additionally, we develop a Feedback-Update training strategy that refines the slicing decisions using feedback from the downstream artificial neural network (ANN). Extensive experiments demonstrate that our method yields significant performance improvements in event-based object tracking and recognition. Notably, SpikeSlicer provides a brand-new SNN-ANN cooperation paradigm, where the SNN acts as an efficient, low-energy data processor to assist the ANN in improving downstream performance, injecting new perspectives and potential avenues of exploration.

comment: Accepted to NeurIPS 2024

☆ Visual Prompting in LLMs for Enhancing Emotion Recognition EMNLP2024

Qixuan Zhang, Zhifeng Wang, Dylan Zhang, Wenjia Niu, Sabrina Caldwell, Tom Gedeon, Yang Liu, Zhenyue Qin

Vision Large Language Models (VLLMs) are transforming the intersection of computer vision and natural language processing. Nonetheless, the potential of using visual prompts for emotion recognition in these models remains largely unexplored and untapped. Traditional methods in VLLMs struggle with spatial localization and often discard valuable global context. To address this problem, we propose a Set-of-Vision prompting (SoV) approach that enhances zero-shot emotion recognition by using spatial information, such as bounding boxes and facial landmarks, to mark targets precisely. SoV improves accuracy in face count and emotion categorization while preserving the enriched image context. Through a battery of experimentation and analysis of recent commercial or open-source VLLMs, we evaluate the SoV model's ability to comprehend facial expressions in natural environments. Our findings demonstrate the effectiveness of integrating spatial visual prompts into VLLMs for improving emotion recognition performance.

comment: Accepted by EMNLP2024 (Main, Long paper)

☆ SCA: Highly Efficient Semantic-Consistent Unrestricted Adversarial Attack

Zihao Pan, Weibin Wu, Yuhang Cao, Zibin Zheng

☆ Key-Grid: Unsupervised 3D Keypoints Detection using Grid Heatmap Features

Chengkai Hou, Zhengrong Xue, Bingyang Zhou, Jinghan Ke, Lin Shao, Huazhe Xu

Detecting 3D keypoints with semantic consistency is widely used in many scenarios such as pose estimation, shape registration and robotics. Currently, most unsupervised 3D keypoint detection methods focus on the rigid-body objects. However, when faced with deformable objects, the keypoints they identify do not preserve semantic consistency well. In this paper, we introduce an innovative unsupervised keypoint detector Key-Grid for both the rigid-body and deformable objects, which is an autoencoder framework. The encoder predicts keypoints and the decoder utilizes the generated keypoints to reconstruct the objects. Unlike previous work, we leverage the identified keypoint in formation to form a 3D grid feature heatmap called grid heatmap, which is used in the decoder section. Grid heatmap is a novel concept that represents the latent variables for grid points sampled uniformly in the 3D cubic space, where these variables are the shortest distance between the grid points and the skeleton connected by keypoint pairs. Meanwhile, we incorporate the information from each layer of the encoder into the decoder section. We conduct an extensive evaluation of Key-Grid on a list of benchmark datasets. Key-Grid achieves the state-of-the-art performance on the semantic consistency and position accuracy of keypoints. Moreover, we demonstrate the robustness of Key-Grid to noise and downsampling. In addition, we achieve SE-(3) invariance of keypoints though generalizing Key-Grid to a SE(3)-invariant backbone.

☆ Efficient Semantic Segmentation via Lightweight Multiple-Information Interaction Network

Yangyang Qiu, Guoan Xu, Guangwei Gao, Zhenhua Guo, Yi Yu, Chia-Wen Lin

Recently, the integration of the local modeling capabilities of Convolutional Neural Networks (CNNs) with the global dependency strengths of Transformers has created a sensation in the semantic segmentation community. However, substantial computational workloads and high hardware memory demands remain major obstacles to their further application in real-time scenarios. In this work, we propose a lightweight multiple-information interaction network for real-time semantic segmentation, called LMIINet, which effectively combines CNNs and Transformers while reducing redundant computations and memory footprint. It features Lightweight Feature Interaction Bottleneck (LFIB) modules comprising efficient convolutions that enhance context integration. Additionally, improvements are made to the Flatten Transformer by enhancing local and global feature interaction to capture detailed semantic information. The incorporation of a combination coefficient learning scheme in both LFIB and Transformer blocks facilitates improved feature interaction. Extensive experiments demonstrate that LMIINet excels in balancing accuracy and efficiency. With only 0.72M parameters and 11.74G FLOPs, LMIINet achieves 72.0% mIoU at 100 FPS on the Cityscapes test set and 69.94% mIoU at 160 FPS on the CamVid test dataset using a single RTX2080Ti GPU.

comment: 10 pages, 6 figures, 9 tables

☆ Capturing complex hand movements and object interactions using machine learning-powered stretchable smart textile gloves

Arvin Tashakori, Zenan Jiang, Amir Servati, Saeid Soltanian, Harishkumar Narayana, Katherine Le, Caroline Nakayama, Chieh-ling Yang, Z. Jane Wang, Janice J. Eng, Peyman Servati

Accurate real-time tracking of dexterous hand movements and interactions has numerous applications in human-computer interaction, metaverse, robotics, and tele-health. Capturing realistic hand movements is challenging because of the large number of articulations and degrees of freedom. Here, we report accurate and dynamic tracking of articulated hand and finger movements using stretchable, washable smart gloves with embedded helical sensor yarns and inertial measurement units. The sensor yarns have a high dynamic range, responding to low 0.005 % to high 155 % strains, and show stability during extensive use and washing cycles. We use multi-stage machine learning to report average joint angle estimation root mean square errors of 1.21 and 1.45 degrees for intra- and inter-subjects cross-validation, respectively, matching accuracy of costly motion capture cameras without occlusion or field of view limitations. We report a data augmentation technique that enhances robustness to noise and variations of sensors. We demonstrate accurate tracking of dexterous hand movements during object interactions, opening new avenues of applications including accurate typing on a mock paper keyboard, recognition of complex dynamic and static gestures adapted from American Sign Language and object identification.

☆ Stochastic Sampling from Deterministic Flow Models ICLR 2025

Saurabh Singh, Ian Fischer

Deterministic flow models, such as rectified flows, offer a general framework for learning a deterministic transport map between two distributions, realized as the vector field for an ordinary differential equation (ODE). However, they are sensitive to model estimation and discretization errors and do not permit different samples conditioned on an intermediate state, limiting their application. We present a general method to turn the underlying ODE of such flow models into a family of stochastic differential equations (SDEs) that have the same marginal distributions. This method permits us to derive families of \emph{stochastic samplers}, for fixed (e.g., previously trained) \emph{deterministic} flow models, that continuously span the spectrum of deterministic and stochastic sampling, given access to the flow field and the score function. Our method provides additional degrees of freedom that help alleviate the issues with the deterministic samplers and empirically outperforms them. We empirically demonstrate advantages of our method on a toy Gaussian setup and on the large scale ImageNet generation task. Further, our family of stochastic samplers provide an additional knob for controlling the diversity of generation, which we qualitatively demonstrate in our experiments.

comment: Submitted to ICLR 2025

☆ Hard Negative Sample Mining for Whole Slide Image Classification MICCAI 2024

Wentao Huang, Xiaoling Hu, Shahira Abousamra, Prateek Prasanna, Chao Chen

Weakly supervised whole slide image (WSI) classification is challenging due to the lack of patch-level labels and high computational costs. State-of-the-art methods use self-supervised patch-wise feature representations for multiple instance learning (MIL). Recently, methods have been proposed to fine-tune the feature representation on the downstream task using pseudo labeling, but mostly focusing on selecting high-quality positive patches. In this paper, we propose to mine hard negative samples during fine-tuning. This allows us to obtain better feature representations and reduce the training cost. Furthermore, we propose a novel patch-wise ranking loss in MIL to better exploit these hard negative samples. Experiments on two public datasets demonstrate the efficacy of these proposed ideas. Our codes are available at https://github.com/winston52/HNM-WSI

comment: 13 pages, 4 figures, accepted by MICCAI 2024

☆ Adapting Segment Anything Model to Melanoma Segmentation in Microscopy Slide Images

Qingyuan Liu, Avideh Zakhor

Melanoma segmentation in Whole Slide Images (WSIs) is useful for prognosis and the measurement of crucial prognostic factors such as Breslow depth and primary invasive tumor size. In this paper, we present a novel approach that uses the Segment Anything Model (SAM) for automatic melanoma segmentation in microscopy slide images. Our method employs an initial semantic segmentation model to generate preliminary segmentation masks that are then used to prompt SAM. We design a dynamic prompting strategy that uses a combination of centroid and grid prompts to achieve optimal coverage of the super high-resolution slide images while maintaining the quality of generated prompts. To optimize for invasive melanoma segmentation, we further refine the prompt generation process by implementing in-situ melanoma detection and low-confidence region filtering. We select Segformer as the initial segmentation model and EfficientSAM as the segment anything model for parameter-efficient fine-tuning. Our experimental results demonstrate that this approach not only surpasses other state-of-the-art melanoma segmentation methods but also significantly outperforms the baseline Segformer by 9.1% in terms of IoU.

☆ Remember and Recall: Associative-Memory-based Trajectory Prediction

Hang Guo, Yuzhen Zhang, Tianci Gao, Junning Su, Pei Lv, Mingliang Xu

Trajectory prediction is a pivotal component of autonomous driving systems, enabling the application of accumulated movement experience to current scenarios. Although most existing methods concentrate on learning continuous representations to gain valuable experience, they often suffer from computational inefficiencies and struggle with unfamiliar situations. To address this issue, we propose the Fragmented-Memory-based Trajectory Prediction (FMTP) model, inspired by the remarkable learning capabilities of humans, particularly their ability to leverage accumulated experience and recall relevant memories in unfamiliar situations. The FMTP model employs discrete representations to enhance computational efficiency by reducing information redundancy while maintaining the flexibility to utilize past experiences. Specifically, we design a learnable memory array by consolidating continuous trajectory representations from the training set using defined quantization operations during the training phase. This approach further eliminates redundant information while preserving essential features in discrete form. Additionally, we develop an advanced reasoning engine based on language models to deeply learn the associative rules among these discrete representations. Our method has been evaluated on various public datasets, including ETH-UCY, inD, SDD, nuScenes, Waymo, and VTL-TP. The extensive experimental results demonstrate that our approach achieves significant performance and extracts more valuable experience from past trajectories to inform the current state.

☆ BadCM: Invisible Backdoor Attack Against Cross-Modal Learning

Zheng Zhang, Xu Yuan, Lei Zhu, Jingkuan Song, Liqiang Nie

Despite remarkable successes in unimodal learning tasks, backdoor attacks against cross-modal learning are still underexplored due to the limited generalization and inferior stealthiness when involving multiple modalities. Notably, since works in this area mainly inherit ideas from unimodal visual attacks, they struggle with dealing with diverse cross-modal attack circumstances and manipulating imperceptible trigger samples, which hinders their practicability in real-world applications. In this paper, we introduce a novel bilateral backdoor to fill in the missing pieces of the puzzle in the cross-modal backdoor and propose a generalized invisible backdoor framework against cross-modal learning (BadCM). Specifically, a cross-modal mining scheme is developed to capture the modality-invariant components as target poisoning areas, where well-designed trigger patterns injected into these regions can be efficiently recognized by the victim models. This strategy is adapted to different image-text cross-modal models, making our framework available to various attack scenarios. Furthermore, for generating poisoned samples of high stealthiness, we conceive modality-specific generators for visual and linguistic modalities that facilitate hiding explicit trigger patterns in modality-invariant regions. To the best of our knowledge, BadCM is the first invisible backdoor method deliberately designed for diverse cross-modal attacks within one unified framework. Comprehensive experimental evaluations on two typical applications, i.e., cross-modal retrieval and VQA, demonstrate the effectiveness and generalization of our method under multiple kinds of attack scenarios. Moreover, we show that BadCM can robustly evade existing backdoor defenses. Our code is available at https://github.com/xandery-geek/BadCM.

☆ HATFormer: Historic Handwritten Arabic Text Recognition with Transformers

Adrian Chan, Anupam Mijar, Mehreen Saeed, Chau-Wai Wong, Akram Khater

Arabic handwritten text recognition (HTR) is challenging, especially for historical texts, due to diverse writing styles and the intrinsic features of Arabic script. Additionally, Arabic handwriting datasets are smaller compared to English ones, making it difficult to train generalizable Arabic HTR models. To address these challenges, we propose HATFormer, a transformer-based encoder-decoder architecture that builds on a state-of-the-art English HTR model. By leveraging the transformer's attention mechanism, HATFormer captures spatial contextual information to address the intrinsic challenges of Arabic script through differentiating cursive characters, decomposing visual representations, and identifying diacritics. Our customization to historical handwritten Arabic includes an image processor for effective ViT information preprocessing, a text tokenizer for compact Arabic text representation, and a training pipeline that accounts for a limited amount of historic Arabic handwriting data. HATFormer achieves a character error rate (CER) of 8.6% on the largest public historical handwritten Arabic dataset, with a 51% improvement over the best baseline in the literature. HATFormer also attains a comparable CER of 4.2% on the largest private non-historical dataset. Our work demonstrates the feasibility of adapting an English HTR method to a low-resource language with complex, language-specific challenges, contributing to advancements in document digitization, information retrieval, and cultural preservation.

☆ From Pixels to Tokens: Byte-Pair Encoding on Quantized Visual Modalities

Wanpeng Zhang, Zilong Xie, Yicheng Feng, Yijiang Li, Xingrun Xing, Sipeng Zheng, Zongqing Lu

☆ An Evaluation of Large Pre-Trained Models for Gesture Recognition using Synthetic Videos SP

Arun Reddy, Ketul Shah, Corban Rivera, William Paul, Celso M. De Melo, Rama Chellappa

In this work, we explore the possibility of using synthetically generated data for video-based gesture recognition with large pre-trained models. We consider whether these models have sufficiently robust and expressive representation spaces to enable "training-free" classification. Specifically, we utilize various state-of-the-art video encoders to extract features for use in k-nearest neighbors classification, where the training data points are derived from synthetic videos only. We compare these results with another training-free approach -- zero-shot classification using text descriptions of each gesture. In our experiments with the RoCoG-v2 dataset, we find that using synthetic training videos yields significantly lower classification accuracy on real test videos compared to using a relatively small number of real training videos. We also observe that video backbones that were fine-tuned on classification tasks serve as superior feature extractors, and that the choice of fine-tuning data has a substantial impact on k-nearest neighbors performance. Lastly, we find that zero-shot text-based classification performs poorly on the gesture recognition task, as gestures are not easily described through natural language.

comment: Synthetic Data for Artificial Intelligence and Machine Learning: Tools, Techniques, and Applications II (SPIE Defense + Commercial Sensing, 2024)

☆ MDSGen: Fast and Efficient Masked Diffusion Temporal-Aware Transformers for Open-Domain Sound Generation

Trung X. Pham, Tri Ton, Chang D. Yoo

We introduce MDSGen, a novel framework for vision-guided open-domain sound generation optimized for model parameter size, memory consumption, and inference speed. This framework incorporates two key innovations: (1) a redundant video feature removal module that filters out unnecessary visual information, and (2) a temporal-aware masking strategy that leverages temporal context for enhanced audio generation accuracy. In contrast to existing resource-heavy Unet-based models, MDSGen employs denoising masked diffusion transformers, facilitating efficient generation without reliance on pre-trained diffusion models. Evaluated on the benchmark VGGSound dataset, our smallest model (5M parameters) achieves 97.9% alignment accuracy, using 172x fewer parameters, 371% less memory, and offering 36x faster inference than the current 860M-parameter state-of-the-art model (93.9% accuracy). The larger model (131M parameters) reaches nearly 99% accuracy while requiring 6.5x fewer parameters. These results highlight the scalability and effectiveness of our approach.

comment: 21 pages, 16 figures

☆ DMC-Net: Lightweight Dynamic Multi-Scale and Multi-Resolution Convolution Network for Pancreas Segmentation in CT Images

Jin Yang, Daniel S. Marcus, Aristeidis Sotiras

Convolutional neural networks (CNNs) have shown great effectiveness in medical image segmentation. However, they may be limited in modeling large inter-subject variations in organ shapes and sizes and exploiting global long-range contextual information. This is because CNNs typically employ convolutions with fixed-sized local receptive fields and lack the mechanisms to utilize global information. To address these limitations, we developed Dynamic Multi-Resolution Convolution (DMRC) and Dynamic Multi-Scale Convolution (DMSC) modules. Both modules enhance the representation capabilities of single convolutions to capture varying scaled features and global contextual information. This is achieved in the DMRC module by employing a convolutional filter on images with different resolutions and subsequently utilizing dynamic mechanisms to model global inter-dependencies between features. In contrast, the DMSC module extracts features at different scales by employing convolutions with different kernel sizes and utilizing dynamic mechanisms to extract global contextual information. The utilization of convolutions with different kernel sizes in the DMSC module may increase computational complexity. To lessen this burden, we propose to use a lightweight design for convolution layers with a large kernel size. Thus, DMSC and DMRC modules are designed as lightweight drop-in replacements for single convolutions, and they can be easily integrated into general CNN architectures for end-to-end training. The segmentation network was proposed by incorporating our DMSC and DMRC modules into a standard U-Net architecture, termed Dynamic Multi-scale and Multi-resolution Convolution network (DMC-Net). The results demonstrate that our proposed DMSC and DMRC can enhance the representation capabilities of single convolutions and improve segmentation accuracy.

comment: 14 pages, 4 figures

☆ Revealing the Unseen: Guiding Personalized Diffusion Models to Expose Training Data

Xiaoyu Wu, Jiaru Zhang, Steven Wu

Diffusion Models (DMs) have evolved into advanced image generation tools, especially for few-shot fine-tuning where a pretrained DM is fine-tuned on a small set of images to capture specific styles or objects. Many people upload these personalized checkpoints online, fostering communities such as Civitai and HuggingFace. However, model owners may overlook the potential risks of data leakage by releasing their fine-tuned checkpoints. Moreover, concerns regarding copyright violations arise when unauthorized data is used during fine-tuning. In this paper, we ask: "Can training data be extracted from these fine-tuned DMs shared online?" A successful extraction would present not only data leakage threats but also offer tangible evidence of copyright infringement. To answer this, we propose FineXtract, a framework for extracting fine-tuning data. Our method approximates fine-tuning as a gradual shift in the model's learned distribution -- from the original pretrained DM toward the fine-tuning data. By extrapolating the models before and after fine-tuning, we guide the generation toward high-probability regions within the fine-tuned data distribution. We then apply a clustering algorithm to extract the most probable images from those generated using this extrapolated guidance. Experiments on DMs fine-tuned with datasets such as WikiArt, DreamBooth, and real-world checkpoints posted online validate the effectiveness of our method, extracting approximately 20% of fine-tuning data in most cases, significantly surpassing baseline performance.

comment: Under review

☆ Dynamic Sparse Training versus Dense Training: The Unexpected Winner in Image Corruption Robustness

Boqian Wu, Qiao Xiao, Shunxin Wang, Nicola Strisciuglio, Mykola Pechenizkiy, Maurice van Keulen, Decebal Constantin Mocanu, Elena Mocanu

It is generally perceived that Dynamic Sparse Training opens the door to a new era of scalability and efficiency for artificial neural networks at, perhaps, some costs in accuracy performance for the classification task. At the same time, Dense Training is widely accepted as being the "de facto" approach to train artificial neural networks if one would like to maximize their robustness against image corruption. In this paper, we question this general practice. Consequently, we claim that, contrary to what is commonly thought, the Dynamic Sparse Training methods can consistently outperform Dense Training in terms of robustness accuracy, particularly if the efficiency aspect is not considered as a main objective (i.e., sparsity levels between 10% and up to 50%), without adding (or even reducing) resource cost. We validate our claim on two types of data, images and videos, using several traditional and modern deep learning architectures for computer vision and three widely studied Dynamic Sparse Training algorithms. Our findings reveal a new yet-unknown benefit of Dynamic Sparse Training and open new possibilities in improving deep learning robustness beyond the current state of the art.

☆ PixelShuffler: A Simple Image Translation Through Pixel Rearrangement

Omar Zamzam

Image-to-image translation is a topic in computer vision that has a vast range of use cases ranging from medical image translation, such as converting MRI scans to CT scans or to other MRI contrasts, to image colorization, super-resolution, domain adaptation, and generating photorealistic images from sketches or semantic maps. Image style transfer is also a widely researched application of image-to-image translation, where the goal is to synthesize an image that combines the content of one image with the style of another. Existing state-of-the-art methods often rely on complex neural networks, including diffusion models and language models, to achieve high-quality style transfer, but these methods can be computationally expensive and intricate to implement. In this paper, we propose a novel pixel shuffle method that addresses the image-to-image translation problem generally with a specific demonstrative application in style transfer. The proposed method approaches style transfer by shuffling the pixels of the style image such that the mutual information between the shuffled image and the content image is maximized. This approach inherently preserves the colors of the style image while ensuring that the structural details of the content image are retained in the stylized output. We demonstrate that this simple and straightforward method produces results that are comparable to state-of-the-art techniques, as measured by the Learned Perceptual Image Patch Similarity (LPIPS) loss for content preservation and the Fr\'echet Inception Distance (FID) score for style similarity. Our experiments validate that the proposed pixel shuffle method achieves competitive performance with significantly reduced complexity, offering a promising alternative for efficient image style transfer, as well as a promise in usability of the method in general image-to-image translation tasks.

☆ MMP: Towards Robust Multi-Modal Learning with Masked Modality Projection

Niki Nezakati, Md Kaykobad Reza, Ameya Patil, Mashhour Solh, M. Salman Asif

Multimodal learning seeks to combine data from multiple input sources to enhance the performance of different downstream tasks. In real-world scenarios, performance can degrade substantially if some input modalities are missing. Existing methods that can handle missing modalities involve custom training or adaptation steps for each input modality combination. These approaches are either tied to specific modalities or become computationally expensive as the number of input modalities increases. In this paper, we propose Masked Modality Projection (MMP), a method designed to train a single model that is robust to any missing modality scenario. We achieve this by randomly masking a subset of modalities during training and learning to project available input modalities to estimate the tokens for the masked modalities. This approach enables the model to effectively learn to leverage the information from the available modalities to compensate for the missing ones, enhancing missing modality robustness. We conduct a series of experiments with various baseline models and datasets to assess the effectiveness of this strategy. Experiments demonstrate that our approach improves robustness to different missing modality scenarios, outperforming existing methods designed for missing modalities or specific modality combinations.

☆ Fully Automated CTC Detection, Segmentation and Classification for Multi-Channel IF Imaging MICCAI 2024

Evan Schwab, Bharat Annaldas, Nisha Ramesh, Anna Lundberg, Vishal Shelke, Xinran Xu, Cole Gilbertson, Jiyun Byun, Ernest T. Lam

Liquid biopsies (eg., blood draws) offer a less invasive and non-localized alternative to tissue biopsies for monitoring the progression of metastatic breast cancer (mBCa). Immunofluoresence (IF) microscopy is a tool to image and analyze millions of blood cells in a patient sample. By detecting and genetically sequencing circulating tumor cells (CTCs) in the blood, personalized treatment plans are achievable for various cancer subtypes. However, CTCs are rare (about 1 in 2M), making manual CTC detection very difficult. In addition, clinicians rely on quantitative cellular biomarkers to manually classify CTCs. This requires prior tasks of cell detection, segmentation and feature extraction. To assist clinicians, we have developed a fully automated machine learning-based production-level pipeline to efficiently detect, segment and classify CTCs in multi-channel IF images. We achieve over 99% sensitivity and 97% specificity on 9,533 cells from 15 mBCa patients. Our pipeline has been successfully deployed on real mBCa patients, reducing a patient average of 14M detected cells to only 335 CTC candidates for manual review.

comment: Published in MICCAI 2024 MOVI Workshop Conference Proceedings

☆ GABIC: Graph-based Attention Block for Image Compression ICIP 2024

Gabriele Spadaro, Alberto Presta, Enzo Tartaglione, Jhony H. Giraldo, Marco Grangetto, Attilio Fiandrotti

While standardized codecs like JPEG and HEVC-intra represent the industry standard in image compression, neural Learned Image Compression (LIC) codecs represent a promising alternative. In detail, integrating attention mechanisms from Vision Transformers into LIC models has shown improved compression efficiency. However, extra efficiency often comes at the cost of aggregating redundant features. This work proposes a Graph-based Attention Block for Image Compression (GABIC), a method to reduce feature redundancy based on a k-Nearest Neighbors enhanced attention mechanism. Our experiments show that GABIC outperforms comparable methods, particularly at high bit rates, enhancing compression performance.

comment: 10 pages, 5 figures, accepted at ICIP 2024

☆ SymmetricDiffusers: Learning Discrete Diffusion on Finite Symmetric Groups

Yongxing Zhang, Donglin Yang, Renjie Liao

Finite symmetric groups $S_n$ are essential in fields such as combinatorics, physics, and chemistry. However, learning a probability distribution over $S_n$ poses significant challenges due to its intractable size and discrete nature. In this paper, we introduce SymmetricDiffusers, a novel discrete diffusion model that simplifies the task of learning a complicated distribution over $S_n$ by decomposing it into learning simpler transitions of the reverse diffusion using deep neural networks. We identify the riffle shuffle as an effective forward transition and provide empirical guidelines for selecting the diffusion length based on the theory of random walks on finite groups. Additionally, we propose a generalized Plackett-Luce (PL) distribution for the reverse transition, which is provably more expressive than the PL distribution. We further introduce a theoretically grounded "denoising schedule" to improve sampling and learning efficiency. Extensive experiments show that our model achieves state-of-the-art or comparable performances on solving tasks including sorting 4-digit MNIST images, jigsaw puzzles, and traveling salesman problems. Our code is released at https://github.com/NickZhang53/SymmetricDiffusers.

☆ RSA: Resolving Scale Ambiguities in Monocular Depth Estimators through Language Descriptions

Ziyao Zeng, Yangchao Wu, Hyoungseob Park, Daniel Wang, Fengyu Yang, Stefano Soatto, Dong Lao, Byung-Woo Hong, Alex Wong

We propose a method for metric-scale monocular depth estimation. Inferring depth from a single image is an ill-posed problem due to the loss of scale from perspective projection during the image formation process. Any scale chosen is a bias, typically stemming from training on a dataset; hence, existing works have instead opted to use relative (normalized, inverse) depth. Our goal is to recover metric-scaled depth maps through a linear transformation. The crux of our method lies in the observation that certain objects (e.g., cars, trees, street signs) are typically found or associated with certain types of scenes (e.g., outdoor). We explore whether language descriptions can be used to transform relative depth predictions to those in metric scale. Our method, RSA, takes as input a text caption describing objects present in an image and outputs the parameters of a linear transformation which can be applied globally to a relative depth map to yield metric-scaled depth predictions. We demonstrate our method on recent general-purpose monocular depth models on indoors (NYUv2) and outdoors (KITTI). When trained on multiple datasets, RSA can serve as a general alignment module in zero-shot settings. Our method improves over common practices in aligning relative to metric depth and results in predictions that are comparable to an upper bound of fitting relative depth to ground truth via a linear transformation.

☆ AirLetters: An Open Video Dataset of Characters Drawn in the Air ECCV'24

Rishit Dagli, Guillaume Berger, Joanna Materzynska, Ingo Bax, Roland Memisevic

We introduce AirLetters, a new video dataset consisting of real-world videos of human-generated, articulated motions. Specifically, our dataset requires a vision model to predict letters that humans draw in the air. Unlike existing video datasets, accurate classification predictions for AirLetters rely critically on discerning motion patterns and on integrating long-range information in the video over time. An extensive evaluation of state-of-the-art image and video understanding models on AirLetters shows that these methods perform poorly and fall far behind a human baseline. Our work shows that, despite recent progress in end-to-end video understanding, accurate representations of complex articulated motions -- a task that is trivial for humans -- remains an open problem for end-to-end learning.

comment: ECCV'24, HANDS workshop

☆ Task-Decoupled Image Inpainting Framework for Class-specific Object Remover

Changsuk Oh, H. Jin Kim

Object removal refers to the process of erasing designated objects from an image while preserving the overall appearance. Existing works on object removal erase removal targets using image inpainting networks. However, image inpainting networks often generate unsatisfactory removal results. In this work, we find that the current training approach which encourages a single image inpainting model to handle both object removal and restoration tasks is one of the reasons behind such unsatisfactory result. Based on this finding, we propose a task-decoupled image inpainting framework which generates two separate inpainting models: an object restorer for object restoration tasks and an object remover for object removal tasks. We train the object restorer with the masks that partially cover the removal targets. Then, the proposed framework makes an object restorer to generate a guidance for training the object remover. Using the proposed framework, we obtain a class-specific object remover which focuses on removing objects of a target class, aiming to better erase target class objects than general object removers. We also introduce a data curation method that encompasses the image selection and mask generation approaches used to produce training data for the proposed class-specific object remover. Using the proposed curation method, we can simulate the scenarios where an object remover is trained on the data with object removal ground truth images. Experiments on multiple datasets show that the proposed class-specific object remover can better remove target class objects than object removers based on image inpainting networks.

☆ Individuation of 3D perceptual units from neurogeometry of binocular cells

Maria Virginia Bolelli, Giovanna Citti, Alessandro Sarti, Steven W. Zucker

We model the functional architecture of the early stages of three-dimensional vision by extending the neurogeometric sub-Riemannian model for stereo-vision introduced in \cite{BCSZ23}. A new framework for correspondence is introduced that integrates a neural-based algorithm to achieve stereo correspondence locally while, simultaneously, organizing the corresponding points into global perceptual units. The result is an effective scene segmentation. We achieve this using harmonic analysis on the sub-Riemannian structure and show, in a comparison against Riemannian distance, that the sub-Riemannian metric is central to the solution.

comment: 30 pages, 13 figures

♻ ☆ CMP: Cooperative Motion Prediction with Multi-Agent Communication

Zehao Wang, Yuping Wang, Zhuoyuan Wu, Hengbo Ma, Zhaowei Li, Hang Qiu, Jiachen Li

The confluence of the advancement of Autonomous Vehicles (AVs) and the maturity of Vehicle-to-Everything (V2X) communication has enabled the capability of cooperative connected and automated vehicles (CAVs). Building on top of cooperative perception, this paper explores the feasibility and effectiveness of cooperative motion prediction. Our method, CMP, takes LiDAR signals as model input to enhance tracking and prediction capabilities. Unlike previous work that focuses separately on either cooperative perception or motion prediction, our framework, to the best of our knowledge, is the first to address the unified problem where CAVs share information in both perception and prediction modules. Incorporated into our design is the unique capability to tolerate realistic V2X bandwidth limitations and transmission delays, while dealing with bulky perception representations. We also propose a prediction aggregation module, which unifies the predictions obtained by different CAVs and generates the final prediction. Through extensive experiments and ablation studies on the OPV2V and V2V4Real datasets, we demonstrate the effectiveness of our method in cooperative perception, tracking, and motion prediction. In particular, CMP reduces the average prediction error by 16.4\% with fewer missing detections compared with the no cooperation setting and by 12.3\% compared with the strongest baseline. Our work marks a significant step forward in the cooperative capabilities of CAVs, showcasing enhanced performance in complex scenarios. The code can be found on the project website: https://cmp-cooperative-prediction.github.io/.

comment: Project website: https://cmp-cooperative-prediction.github.io/

♻ ☆ NVDS+: Towards Efficient and Versatile Neural Stabilizer for Video Depth Estimation ICCV 2023

Yiran Wang, Min Shi, Jiaqi Li, Chaoyi Hong, Zihao Huang, Juewen Peng, Zhiguo Cao, Jianming Zhang, Ke Xian, Guosheng Lin

Video depth estimation aims to infer temporally consistent depth. One approach is to finetune a single-image model on each video with geometry constraints, which proves inefficient and lacks robustness. An alternative is learning to enforce consistency from data, which requires well-designed models and sufficient video depth data. To address both challenges, we introduce NVDS+ that stabilizes inconsistent depth estimated by various single-image models in a plug-and-play manner. We also elaborate a large-scale Video Depth in the Wild (VDW) dataset, which contains 14,203 videos with over two million frames, making it the largest natural-scene video depth dataset. Additionally, a bidirectional inference strategy is designed to improve consistency by adaptively fusing forward and backward predictions. We instantiate a model family ranging from small to large scales for different applications. The method is evaluated on VDW dataset and three public benchmarks. To further prove the versatility, we extend NVDS+ to video semantic segmentation and several downstream applications like bokeh rendering, novel view synthesis, and 3D reconstruction. Experimental results show that our method achieves significant improvements in consistency, accuracy, and efficiency. Our work serves as a solid baseline and data foundation for learning-based video depth estimation. Code and dataset are available at: https://github.com/RaymondWang987/NVDS

comment: V1/V2: ICCV 2023 accepted; V3: the journal extension accepted by IEEE TPAMI 2024

♻ ☆ SwapAnything: Enabling Arbitrary Object Swapping in Personalized Visual Editing ECCV 2024

Jing Gu, Nanxuan Zhao, Wei Xiong, Qing Liu, Zhifei Zhang, He Zhang, Jianming Zhang, HyunJoon Jung, Yilin Wang, Xin Eric Wang

Effective editing of personal content holds a pivotal role in enabling individuals to express their creativity, weaving captivating narratives within their visual stories, and elevate the overall quality and impact of their visual content. Therefore, in this work, we introduce SwapAnything, a novel framework that can swap any objects in an image with personalized concepts given by the reference, while keeping the context unchanged. Compared with existing methods for personalized subject swapping, SwapAnything has three unique advantages: (1) precise control of arbitrary objects and parts rather than the main subject, (2) more faithful preservation of context pixels, (3) better adaptation of the personalized concept to the image. First, we propose targeted variable swapping to apply region control over latent feature maps and swap masked variables for faithful context preservation and initial semantic concept swapping. Then, we introduce appearance adaptation, to seamlessly adapt the semantic concept into the original image in terms of target location, shape, style, and content during the image generation process. Extensive results on both human and automatic evaluation demonstrate significant improvements of our approach over baseline methods on personalized swapping. Furthermore, SwapAnything shows its precise and faithful swapping abilities across single object, multiple objects, partial object, and cross-domain swapping tasks. SwapAnything also achieves great performance on text-based swapping and tasks beyond swapping such as object insertion.

comment: ECCV 2024, 23 pages, 14 figures, 3 tables

♻ ☆ Towards Foundation Models and Few-Shot Parameter-Efficient Fine-Tuning for Volumetric Organ Segmentation MICCAI

Julio Silva-Rodríguez, Jose Dolz, Ismail Ben Ayed

The recent popularity of foundation models and the pre-train-and-adapt paradigm, where a large-scale model is transferred to downstream tasks, is gaining attention for volumetric medical image segmentation. However, current transfer learning strategies devoted to full fine-tuning for transfer learning may require significant resources and yield sub-optimal results when the labeled data of the target task is scarce. This makes its applicability in real clinical settings challenging since these institutions are usually constrained on data and computational resources to develop proprietary solutions. To address this challenge, we formalize Few-Shot Efficient Fine-Tuning (FSEFT), a novel and realistic scenario for adapting medical image segmentation foundation models. This setting considers the key role of both data- and parameter- efficiency during adaptation. Building on a foundation model pre-trained on open-access CT organ segmentation sources, we propose leveraging Parameter-Efficient Fine-Tuning and black-box Adapters to address such challenges. Furthermore, novel efficient adaptation methodologies are introduced in this work, which include Spatial black-box Adapters that are more appropriate for dense prediction tasks and constrained transductive inference, leveraging task-specific prior knowledge. Our comprehensive transfer learning experiments confirm the suitability of foundation models in medical image segmentation and unveil the limitations of popular fine-tuning strategies in few-shot scenarios.

comment: Journal Extension of MICCAI - MedAGI Workshop 2023. Code in https://github.com/jusiro/fewshot-finetuning

♻ ☆ Autoregressive Pre-Training on Pixels and Texts EMNLP 2024

Yekun Chai, Qingyi Liu, Jingwu Xiao, Shuohuan Wang, Yu Sun, Hua Wu

comment: EMNLP 2024

♻ ☆ VideoPhy: Evaluating Physical Commonsense for Video Generation

Hritik Bansal, Zongyu Lin, Tianyi Xie, Zeshun Zong, Michal Yarom, Yonatan Bitton, Chenfanfu Jiang, Yizhou Sun, Kai-Wei Chang, Aditya Grover

Recent advances in internet-scale video data pretraining have led to the development of text-to-video generative models that can create high-quality videos across a broad range of visual concepts, synthesize realistic motions and render complex objects. Hence, these generative models have the potential to become general-purpose simulators of the physical world. However, it is unclear how far we are from this goal with the existing text-to-video generative models. To this end, we present VideoPhy, a benchmark designed to assess whether the generated videos follow physical commonsense for real-world activities (e.g. marbles will roll down when placed on a slanted surface). Specifically, we curate diverse prompts that involve interactions between various material types in the physical world (e.g., solid-solid, solid-fluid, fluid-fluid). We then generate videos conditioned on these captions from diverse state-of-the-art text-to-video generative models, including open models (e.g., CogVideoX) and closed models (e.g., Lumiere, Dream Machine). Our human evaluation reveals that the existing models severely lack the ability to generate videos adhering to the given text prompts, while also lack physical commonsense. Specifically, the best performing model, CogVideoX-5B, generates videos that adhere to the caption and physical laws for 39.6% of the instances. VideoPhy thus highlights that the video generative models are far from accurately simulating the physical world. Finally, we propose an auto-evaluator, VideoCon-Physics, to assess the performance reliably for the newly released models.

comment: 43 pages, 29 figures, 12 tables. Added CogVideo and Dream Machine in v2

♻ ☆ Generalizing Medical Image Representations via Quaternion Wavelet Networks

Luigi Sigillo, Eleonora Grassucci, Aurelio Uncini, Danilo Comminiello

Neural network generalizability is becoming a broad research field due to the increasing availability of datasets from different sources and for various tasks. This issue is even wider when processing medical data, where a lack of methodological standards causes large variations being provided by different imaging centers or acquired with various devices and cofactors. To overcome these limitations, we introduce a novel, generalizable, data- and task-agnostic framework able to extract salient features from medical images. The proposed quaternion wavelet network (QUAVE) can be easily integrated with any pre-existing medical image analysis or synthesis task, and it can be involved with real, quaternion, or hypercomplex-valued models, generalizing their adoption to single-channel data. QUAVE first extracts different sub-bands through the quaternion wavelet transform, resulting in both low-frequency/approximation bands and high-frequency/fine-grained features. Then, it weighs the most representative set of sub-bands to be involved as input to any other neural model for image processing, replacing standard data samples. We conduct an extensive experimental evaluation comprising different datasets, diverse image analysis, and synthesis tasks including reconstruction, segmentation, and modality translation. We also evaluate QUAVE in combination with both real and quaternion-valued models. Results demonstrate the effectiveness and the generalizability of the proposed framework that improves network performance while being flexible to be adopted in manifold scenarios and robust to domain shifts. The full code is available at: https://github.com/ispamm/QWT.

comment: This paper is currently under review

♻ ☆ Evaluating Perceptual Distance Models by Fitting Binomial Distributions to Two-Alternative Forced Choice Data

Alexander Hepburn, Raul Santos-Rodriguez, Javier Portilla

The two-alternative forced choice (2AFC) experimental method is popular in the visual perception literature, where practitioners aim to understand how human observers perceive distances within triplets made of a reference image and two distorted versions. In the past, this had been conducted in controlled environments, with triplets sharing images, so it was possible to rank the perceived quality. This ranking would then be used to evaluate perceptual distance models against the experimental data. Recently, crowd-sourced perceptual datasets have emerged, with no images shared between triplets, making ranking infeasible. Evaluating perceptual distance models using this data reduces the judgements on a triplet to a binary decision, namely, whether the distance model agrees with the human decision - which is suboptimal and prone to misleading conclusions. Instead, we statistically model the underlying decision-making process during 2AFC experiments using a binomial distribution. Having enough empirical data, we estimate a smooth and consistent distribution of the judgements on the reference-distorted distance plane, according to each distance model. By applying maximum likelihood, we estimate the parameter of the local binomial distribution, and a global measurement of the expected log-likelihood of the measured responses. We calculate meaningful and well-founded metrics for the distance model, beyond the mere prediction accuracy as percentage agreement, even with variable numbers of judgements per triplet -- key advantages over both classical and neural network methods.

♻ ☆ Context and Geometry Aware Voxel Transformer for Semantic Scene Completion NIPS 2024

Zhu Yu, Runmin Zhang, Jiacheng Ying, Junchen Yu, Xiaohai Hu, Lun Luo, Si-Yuan Cao, Hui-Liang Shen

comment: NIPS 2024 Spotlight

♻ ☆ Leopard: A Vision Language Model For Text-Rich Multi-Image Tasks

Mengzhao Jia, Wenhao Yu, Kaixin Ma, Tianqing Fang, Zhihan Zhang, Siru Ouyang, Hongming Zhang, Meng Jiang, Dong Yu

comment: Our code is available at https://github.com/Jill0001/Leopard

♻ ☆ Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution

comment: Code is available at https://github.com/QwenLM/Qwen2-VL. arXiv admin note: text overlap with arXiv:2408.15262 by other authors

♻ ☆ BinaryDM: Accurate Weight Binarization for Efficient Diffusion Models

Xingyu Zheng, Xianglong Liu, Haotong Qin, Xudong Ma, Mingyuan Zhang, Haojie Hao, Jiakai Wang, Zixiang Zhao, Jinyang Guo, Michele Magno

With the advancement of diffusion models (DMs) and the substantially increased computational requirements, quantization emerges as a practical solution to obtain compact and efficient low-bit DMs. However, the highly discrete representation leads to severe accuracy degradation, hindering the quantization of diffusion models to ultra-low bit-widths. This paper proposes a novel weight binarization approach for DMs, namely BinaryDM, pushing binarized DMs to be accurate and efficient by improving the representation and optimization. From the representation perspective, we present an Evolvable-Basis Binarizer (EBB) to enable a smooth evolution of DMs from full-precision to accurately binarized. EBB enhances information representation in the initial stage through the flexible combination of multiple binary bases and applies regularization to evolve into efficient single-basis binarization. The evolution only occurs in the head and tail of the DM architecture to retain the stability of training. From the optimization perspective, a Low-rank Representation Mimicking (LRM) is applied to assist the optimization of binarized DMs. The LRM mimics the representations of full-precision DMs in low-rank space, alleviating the direction ambiguity of the optimization process caused by fine-grained alignment. Comprehensive experiments demonstrate that BinaryDM achieves significant accuracy and efficiency gains compared to SOTA quantization methods of DMs under ultra-low bit-widths. With 1-bit weight and 4-bit activation (W1A4), BinaryDM achieves as low as 7.74 FID and saves the performance from collapse (baseline FID 10.87). As the first binarization method for diffusion models, W1A4 BinaryDM achieves impressive 15.2x OPs and 29.2x model size savings, showcasing its substantial potential for edge deployment.

comment: The code is available at https://github.com/Xingyu-Zheng/BinaryDM

♻ ☆ Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training NeurIPS 2024

Haoran He, Chenjia Bai, Ling Pan, Weinan Zhang, Bin Zhao, Xuelong Li

comment: Accepted by NeurIPS 2024. 24 pages

♻ ☆ MARVIS: Motion & Geometry Aware Real and Virtual Image Segmentation

Jiayi Wu, Xiaomin Lin, Shahriar Negahdaripour, Cornelia Fermüller, Yiannis Aloimonos

Tasks such as autonomous navigation, 3D reconstruction, and object recognition near the water surfaces are crucial in marine robotics applications. However, challenges arise due to dynamic disturbances, e.g., light reflections and refraction from the random air-water interface, irregular liquid flow, and similar factors, which can lead to potential failures in perception and navigation systems. Traditional computer vision algorithms struggle to differentiate between real and virtual image regions, significantly complicating tasks. A virtual image region is an apparent representation formed by the redirection of light rays, typically through reflection or refraction, creating the illusion of an object's presence without its actual physical location. This work proposes a novel approach for segmentation on real and virtual image regions, exploiting synthetic images combined with domain-invariant information, a Motion Entropy Kernel, and Epipolar Geometric Consistency. Our segmentation network does not need to be re-trained if the domain changes. We show this by deploying the same segmentation network in two different domains: simulation and the real world. By creating realistic synthetic images that mimic the complexities of the water surface, we provide fine-grained training data for our network (MARVIS) to discern between real and virtual images effectively. By motion & geometry-aware design choices and through comprehensive experimental analysis, we achieve state-of-the-art real-virtual image segmentation performance in unseen real world domain, achieving an IoU over 78% and a F1-Score over 86% while ensuring a small computational footprint. MARVIS offers over 43 FPS (8 FPS) inference rates on a single GPU (CPU core). Our code and dataset are available here https://github.com/jiayi-wu-umd/MARVIS.

♻ ☆ Physics-Regularized Multi-Modal Image Assimilation for Brain Tumor Localization NeurIPS 2024

Michal Balcerak, Tamaz Amiranashvili, Andreas Wagner, Jonas Weidner, Petr Karnakov, Johannes C. Paetzold, Ivan Ezhov, Petros Koumoutsakos, Benedikt Wiestler, Bjoern Menze

Physical models in the form of partial differential equations represent an important prior for many under-constrained problems. One example is tumor treatment planning, which heavily depends on accurate estimates of the spatial distribution of tumor cells in a patient's anatomy. Medical imaging scans can identify the bulk of the tumor, but they cannot reveal its full spatial distribution. Tumor cells at low concentrations remain undetectable, for example, in the most frequent type of primary brain tumors, glioblastoma. Deep-learning-based approaches fail to estimate the complete tumor cell distribution due to a lack of reliable training data. Most existing works therefore rely on physics-based simulations to match observed tumors, providing anatomically and physiologically plausible estimations. However, these approaches struggle with complex and unknown initial conditions and are limited by overly rigid physical models. In this work, we present a novel method that balances data-driven and physics-based cost functions. In particular, we propose a unique discretization scheme that quantifies the adherence of our learned spatiotemporal tumor and brain tissue distributions to their corresponding growth and elasticity equations. This quantification, serving as a regularization term rather than a hard constraint, enables greater flexibility and proficiency in assimilating patient data than existing models. We demonstrate improved coverage of tumor recurrence areas compared to existing techniques on real-world data from a cohort of patients. The method holds the potential to enhance clinical adoption of model-driven treatment planning for glioblastoma.

comment: Accepted to NeurIPS 2024

♻ ☆ Source-Free Domain Adaptation Guided by Vision and Vision-Language Pre-Training ICCV

Wenyu Zhang, Li Shen, Chuan-Sheng Foo

Source-free domain adaptation (SFDA) aims to adapt a source model trained on a fully-labeled source domain to a related but unlabeled target domain. While the source model is a key avenue for acquiring target pseudolabels, the generated pseudolabels may exhibit source bias. In the conventional SFDA pipeline, a large data (e.g. ImageNet) pre-trained feature extractor is used to initialize the source model at the start of source training, and subsequently discarded. Despite having diverse features important for generalization, the pre-trained feature extractor can overfit to the source data distribution during source training and forget relevant target domain knowledge. Rather than discarding this valuable knowledge, we introduce an integrated framework to incorporate pre-trained networks into the target adaptation process. The proposed framework is flexible and allows us to plug modern pre-trained networks into the adaptation process to leverage their stronger representation learning capabilities. For adaptation, we propose the Co-learn algorithm to improve target pseudolabel quality collaboratively through the source model and a pre-trained feature extractor. Building on the recent success of the vision-language model CLIP in zero-shot image recognition, we present an extension Co-learn++ to further incorporate CLIP's zero-shot classification decisions. We evaluate on 4 benchmark datasets and include more challenging scenarios such as open-set, partial-set and open-partial SFDA. Experimental results demonstrate that our proposed strategy improves adaptation performance and can be successfully integrated with existing SFDA methods. Project code is available at https://github.com/zwenyu/colearn-plus.

comment: Extension of ICCV paper arXiv:2212.07585; Published at IJCV

♻ ☆ SurgPointTransformer: Vertebrae Shape Completion with RGB-D Data

Aidana Massalimova, Florentin Liebmann, Sascha Jecklin, Fabio Carrillo, Farshad Mazda, Philipp Fürnstahl

State-of-the-art computer- and robot-assisted surgery systems heavily depend on intraoperative imaging technologies such as CT and fluoroscopy to generate detailed 3D visualization of the patient's anatomy. While imaging techniques are highly accurate, they are based on ionizing radiation and expose patients and clinicians. This study introduces an alternative, radiation-free approach for reconstructing the 3D spine anatomy using RGB-D data. Drawing inspiration from the 3D "mental map" that surgeons form during surgeries, we introduce SurgPointTransformer, a shape completion approach for surgical applications that can accurately reconstruct the unexposed spine regions from sparse observations of the exposed surface. Our method involves two main steps: segmentation and shape completion. The segmentation step includes spinal column localization and segmentation, followed by vertebra-wise segmentation. The segmented vertebra point clouds are then subjected to SurgPointTransformer, which leverages an attention mechanism to learn patterns between visible surface features and the underlying anatomy. For evaluation, we utilize an ex-vivo dataset of nine specimens. Their CT data is used to establish ground truth data that were used to compare to the outputs of our methods. Our method significantly outperforms the state-of-the-art baselines, achieving an average Chamfer Distance of 5.39, an F-Score of 0.85, an Earth Mover's Distance of 0.011, and a Signal-to-Noise Ratio of 22.90 dB. This study demonstrates the potential of our reconstruction method for 3D vertebral shape completion. It enables 3D reconstruction of the entire lumbar spine and surgical guidance without ionizing radiation or invasive imaging. Our work contributes to computer-aided and robot-assisted surgery, advancing the perception and intelligence of these systems.

♻ ☆ TransRef: Multi-Scale Reference Embedding Transformer for Reference-Guided Image Inpainting

Taorong Liu, Liang Liao, Delin Chen, Jing Xiao, Zheng Wang, Chia-Wen Lin, Shin'ichi Satoh

Image inpainting for completing complicated semantic environments and diverse hole patterns of corrupted images is challenging even for state-of-the-art learning-based inpainting methods trained on large-scale data. A reference image capturing the same scene of a corrupted image offers informative guidance for completing the corrupted image as it shares similar texture and structure priors to that of the holes of the corrupted image. In this work, we propose a transformer-based encoder-decoder network, named TransRef, for reference-guided image inpainting. Specifically, the guidance is conducted progressively through a reference embedding procedure, in which the referencing features are subsequently aligned and fused with the features of the corrupted image. For precise utilization of the reference features for guidance, a reference-patch alignment (Ref-PA) module is proposed to align the patch features of the reference and corrupted images and harmonize their style differences, while a reference-patch transformer (Ref-PT) module is proposed to refine the embedded reference feature. Moreover, to facilitate the research of reference-guided image restoration tasks, we construct a publicly accessible benchmark dataset containing 50K pairs of input and reference images. Both quantitative and qualitative evaluations demonstrate the efficacy of the reference information and the proposed method over the state-of-the-art methods in completing complex holes. Code and dataset can be accessed at https://github.com/Cameltr/TransRef.

comment: Under review

♻ ☆ Conditional Image Synthesis with Diffusion Models: A Survey

Zheyuan Zhan, Defang Chen, Jian-Ping Mei, Zhenghe Zhao, Jiawei Chen, Chun Chen, Siwei Lyu, Can Wang

Conditional image synthesis based on user-specified requirements is a key component in creating complex visual content. In recent years, diffusion-based generative modeling has become a highly effective way for conditional image synthesis, leading to exponential growth in the literature. However, the complexity of diffusion-based modeling, the wide range of image synthesis tasks, and the diversity of conditioning mechanisms present significant challenges for researchers to keep up with rapid developments and understand the core concepts on this topic. In this survey, we categorize existing works based on how conditions are integrated into the two fundamental components of diffusion-based modeling, i.e., the denoising network and the sampling process. We specifically highlight the underlying principles, advantages, and potential challenges of various conditioning approaches in the training, re-purposing, and specialization stages to construct a desired denoising network. We also summarize six mainstream conditioning mechanisms in the essential sampling process. All discussions are centered around popular applications. Finally, we pinpoint some critical yet still open problems to be solved in the future and suggest some possible solutions. Our reviewed works are itemized at https://github.com/zju-pi/Awesome-Conditional-Diffusion-Models.

♻ ☆ Forecasting Disease Progression with Parallel Hyperplanes in Longitudinal Retinal OCT MICCAI 2024

Arunava Chakravarty, Taha Emre, Dmitrii Lachinov, Antoine Rivail, Hendrik Scholl, Lars Fritsche, Sobha Sivaprasad, Daniel Rueckert, Andrew Lotery, Ursula Schmidt-Erfurth, Hrvoje Bogunović

Predicting future disease progression risk from medical images is challenging due to patient heterogeneity, and subtle or unknown imaging biomarkers. Moreover, deep learning (DL) methods for survival analysis are susceptible to image domain shifts across scanners. We tackle these issues in the task of predicting late dry Age-related Macular Degeneration (dAMD) onset from retinal OCT scans. We propose a novel DL method for survival prediction to jointly predict from the current scan a risk score, inversely related to time-to-conversion, and the probability of conversion within a time interval $t$. It uses a family of parallel hyperplanes generated by parameterizing the bias term as a function of $t$. In addition, we develop unsupervised losses based on intra-subject image pairs to ensure that risk scores increase over time and that future conversion predictions are consistent with AMD stage prediction using actual scans of future visits. Such losses enable data-efficient fine-tuning of the trained model on new unlabeled datasets acquired with a different scanner. Extensive evaluation on two large datasets acquired with different scanners resulted in a mean AUROCs of 0.82 for Dataset-1 and 0.83 for Dataset-2, across prediction intervals of 6,12 and 24 months.

comment: accepted in MICCAI 2024

♻ ☆ Eliciting In-Context Learning in Vision-Language Models for Videos Through Curated Data Distributional Properties EMNLP 2024

Keunwoo Peter Yu, Zheyuan Zhang, Fengyuan Hu, Shane Storks, Joyce Chai

comment: 16 pages, LaTeX; Accepted to EMNLP 2024 Main

♻ ☆ Multimodal Self-Instruct: Synthetic Abstract Image and Visual Reasoning Instruction Using Language Model EMNLP-24

Wenqi Zhang, Zhenglin Cheng, Yuanyu He, Mengna Wang, Yongliang Shen, Zeqi Tan, Guiyang Hou, Mingqian He, Yanna Ma, Weiming Lu, Yueting Zhuang

Although most current large multimodal models (LMMs) can already understand photos of natural scenes and portraits, their understanding of abstract images, e.g., charts, maps, or layouts, and visual reasoning capabilities remains quite rudimentary. They often struggle with simple daily tasks, such as reading time from a clock, understanding a flowchart, or planning a route using a road map. In light of this, we design a multi-modal self-instruct, utilizing large language models and their code capabilities to synthesize massive abstract images and visual reasoning instructions across daily scenarios. Our strategy effortlessly creates a multimodal benchmark with 11,193 instructions for eight visual scenarios: charts, tables, simulated maps, dashboards, flowcharts, relation graphs, floor plans, and visual puzzles. \textbf{This benchmark, constructed with simple lines and geometric elements, exposes the shortcomings of most advanced LMMs} like Claude-3.5-Sonnet and GPT-4o in abstract image understanding, spatial relations reasoning, and visual element induction. Besides, to verify the quality of our synthetic data, we fine-tune an LMM using 62,476 synthetic chart, table and road map instructions. The results demonstrate improved chart understanding and map navigation performance, and also demonstrate potential benefits for other visual reasoning tasks. Our code is available at: \url{https://github.com/zwq2018/Multi-modal-Self-instruct}.

comment: The paper is accepted by EMNLP-24. Code: https://github.com/zwq2018/Multi-modal-Self-instruct dataset: https://huggingface.co/datasets/zwq2018/Multi-modal-Self-instruct Leaderboard: https://multi-modal-self-instruct.github.io/

♻ ☆ SRIF: Semantic Shape Registration Empowered by Diffusion-based Image Morphing and Flow Estimation SIGGRAPH

Mingze Sun, Chen Guo, Puhua Jiang, Shiwei Mao, Yurun Chen, Ruqi Huang

In this paper, we propose SRIF, a novel Semantic shape Registration framework based on diffusion-based Image morphing and Flow estimation. More concretely, given a pair of extrinsically aligned shapes, we first render them from multi-views, and then utilize an image interpolation framework based on diffusion models to generate sequences of intermediate images between them. The images are later fed into a dynamic 3D Gaussian splatting framework, with which we reconstruct and post-process for intermediate point clouds respecting the image morphing processing. In the end, tailored for the above, we propose a novel registration module to estimate continuous normalizing flow, which deforms source shape consistently towards the target, with intermediate point clouds as weak guidance. Our key insight is to leverage large vision models (LVMs) to associate shapes and therefore obtain much richer semantic information on the relationship between shapes than the ad-hoc feature extraction and alignment. As a consequence, SRIF achieves high-quality dense correspondences on challenging shape pairs, but also delivers smooth, semantically meaningful interpolation in between. Empirical evidence justifies the effectiveness and superiority of our method as well as specific design choices. The code is released at https://github.com/rqhuang88/SRIF.

comment: Accepted as a conference paper of SIGGRAPH Asia 2024

♻ ☆ Generalizable Human Gaussians from Single-View Image

Jinnan Chen, Chen Li, Jianfeng Zhang, Lingting Zhu, Buzhen Huang, Hanlin Chen, Gim Hee Lee

In this work, we tackle the task of learning generalizable 3D human Gaussians from a single image. The main challenge for this task is to recover detailed geometry and appearance, especially for the unobserved regions. To this end, we propose single-view generalizable Human Gaussian model (HGM), a diffusion-guided framework for 3D human modeling from a single image. We design a diffusion-based coarse-to-fine pipeline, where the diffusion model is adapted to refine novel-view images rendered from a coarse human Gaussian model. The refined images are then used together with the input image to learn a refined human Gaussian model. Although effective in hallucinating the unobserved views, the approach may generate unrealistic human pose and shapes due to the lack of supervision. We circumvent this problem by further encoding the geometric priors from SMPL model. Specifically, we propagate geometric features from SMPL volume to the predicted Gaussians via sparse convolution and attention mechanism. We validate our approach on publicly available datasets and demonstrate that it significantly surpasses state-of-the-art methods in terms of PSNR and SSIM. Additionally, our method exhibits strong generalization for in-the-wild images.

comment: https://jinnan-chen.github.io/projects/HGM/

♻ ☆ Releasing the Parameter Latency of Neural Representation for High-Efficiency Video Compression

Gai Zhang, Xinfeng Zhang, Lv Tang, Yue Li, Kai Zhang, Li Zhang

For decades, video compression technology has been a prominent research area. Traditional hybrid video compression framework and end-to-end frameworks continue to explore various intra- and inter-frame reference and prediction strategies based on discrete transforms and deep learning techniques. However, the emerging implicit neural representation (INR) technique models entire videos as basic units, automatically capturing intra-frame and inter-frame correlations and obtaining promising performance. INR uses a compact neural network to store video information in network parameters, effectively eliminating spatial and temporal redundancy in the original video. However, in this paper, our exploration and verification reveal that current INR video compression methods do not fully exploit their potential to preserve information. We investigate the potential of enhancing network parameter storage through parameter reuse. By deepening the network, we designed a feasible INR parameter reuse scheme to further improve compression performance. Extensive experimental results show that our method significantly enhances the rate-distortion performance of INR video compression.

♻ ☆ LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via a Hybrid Architecture

Xidong Wang, Dingjie Song, Shunian Chen, Chen Zhang, Benyou Wang

comment: 20 pages, 9 figures, 9 tables

♻ ☆ EUFCC-CIR: a Composed Image Retrieval Dataset for GLAM Collections ECCV

Francesc Net, Lluis Gomez

The intersection of Artificial Intelligence and Digital Humanities enables researchers to explore cultural heritage collections with greater depth and scale. In this paper, we present EUFCC-CIR, a dataset designed for Composed Image Retrieval (CIR) within Galleries, Libraries, Archives, and Museums (GLAM) collections. Our dataset is built on top of the EUFCC-340K image labeling dataset and contains over 180K annotated CIR triplets. Each triplet is composed of a multi-modal query (an input image plus a short text describing the desired attribute manipulations) and a set of relevant target images. The EUFCC-CIR dataset fills an existing gap in CIR-specific resources for Digital Humanities. We demonstrate the value of the EUFCC-CIR dataset by highlighting its unique qualities in comparison to other existing CIR datasets and evaluating the performance of several zero-shot CIR baselines.

comment: ECCV Workshop (AI4DH2024)

♻ ☆ Fake It Until You Break It: On the Adversarial Robustness of AI-generated Image Detectors

Sina Mavali, Jonas Ricker, David Pape, Yash Sharma, Asja Fischer, Lea Schönherr

While generative AI (GenAI) offers countless possibilities for creative and productive tasks, artificially generated media can be misused for fraud, manipulation, scams, misinformation campaigns, and more. To mitigate the risks associated with maliciously generated media, forensic classifiers are employed to identify AI-generated content. However, current forensic classifiers are often not evaluated in practically relevant scenarios, such as the presence of an attacker or when real-world artifacts like social media degradations affect images. In this paper, we evaluate state-of-the-art AI-generated image (AIGI) detectors under different attack scenarios. We demonstrate that forensic classifiers can be effectively attacked in realistic settings, even when the attacker does not have access to the target model and post-processing occurs after the adversarial examples are created, which is standard on social media platforms. These attacks can significantly reduce detection accuracy to the extent that the risks of relying on detectors outweigh their benefits. Finally, we propose a simple defense mechanism to make CLIP-based detectors, which are currently the best-performing detectors, robust against these attacks.

♻ ☆ Optimal Projections for Discriminative Dictionary Learning using the JL-lemma

G. Madhuri, Atul Negi, Kaluri V. Rangarao

Dimensionality reduction-based dictionary learning methods in the literature have often used iterative random projections. The dimensionality of such a random projection matrix is a random number that might not lead to a separable subspace structure in the transformed space. The convergence of such methods highly depends on the initial seed values used. Also, gradient descent-based updates might result in local minima. This paper proposes a constructive approach to derandomize the projection matrix using the Johnson-Lindenstrauss lemma. Rather than reducing dimensionality via random projections, a projection matrix derived from the proposed Modified Supervised PC analysis is used. A heuristic is proposed to decide the data perturbation levels and the dictionary atom's corresponding suitable description length. The projection matrix is derived in a single step, provides maximum feature-label consistency of the transformed space, and preserves the geometry of the original data. The projection matrix thus constructed is proved to be a JL-embedding. Despite confusing classes in the OCR datasets, the dictionary trained in the transformed space generates discriminative sparse coefficients with reduced complexity. Empirical study demonstrates that the proposed method performs well even when the number of classes and dimensionality increase. Experimentation on OCR and face recognition datasets shows better classification performance than other algorithms.

♻ ☆ MOREL: Enhancing Adversarial Robustness through Multi-Objective Representation Learning

Sedjro Salomon Hotegni, Sebastian Peitz

Extensive research has shown that deep neural networks (DNNs) are vulnerable to slight adversarial perturbations$-$small changes to the input data that appear insignificant but cause the model to produce drastically different outputs. In addition to augmenting training data with adversarial examples generated from a specific attack method, most of the current defense strategies necessitate modifying the original model architecture components to improve robustness or performing test-time data purification to handle adversarial attacks. In this work, we demonstrate that strong feature representation learning during training can significantly enhance the original model's robustness. We propose MOREL, a multi-objective feature representation learning approach, encouraging classification models to produce similar features for inputs within the same class, despite perturbations. Our training method involves an embedding space where cosine similarity loss and multi-positive contrastive loss are used to align natural and adversarial features from the model encoder and ensure tight clustering. Concurrently, the classifier is motivated to achieve accurate predictions. Through extensive experiments, we demonstrate that our approach significantly enhances the robustness of DNNs against white-box and black-box adversarial attacks, outperforming other methods that similarly require no architectural changes or test-time data purification. Our code is available at https://github.com/salomonhotegni/MOREL

♻ ☆ Efficient Temporal Extrapolation of Multimodal Large Language Models with Temporal Grounding Bridge EMNLP 2024

Yuxuan Wang, Yueqian Wang, Pengfei Wu, Jianxin Liang, Dongyan Zhao, Yang Liu, Zilong Zheng

comment: To appear at EMNLP 2024

♻ ☆ Text-to-Sticker: Style Tailoring Latent Diffusion Models for Human Expression

Animesh Sinha, Bo Sun, Anmol Kalia, Arantxa Casanova, Elliot Blanchard, David Yan, Winnie Zhang, Tony Nelli, Jiahui Chen, Hardik Shah, Licheng Yu, Mitesh Kumar Singh, Ankit Ramchandani, Maziar Sanjabi, Sonal Gupta, Amy Bearman, Dhruv Mahajan

We introduce Style Tailoring, a recipe to finetune Latent Diffusion Models (LDMs) in a distinct domain with high visual quality, prompt alignment and scene diversity. We choose sticker image generation as the target domain, as the images significantly differ from photorealistic samples typically generated by large-scale LDMs. We start with a competent text-to-image model, like Emu, and show that relying on prompt engineering with a photorealistic model to generate stickers leads to poor prompt alignment and scene diversity. To overcome these drawbacks, we first finetune Emu on millions of sticker-like images collected using weak supervision to elicit diversity. Next, we curate human-in-the-loop (HITL) Alignment and Style datasets from model generations, and finetune to improve prompt alignment and style alignment respectively. Sequential finetuning on these datasets poses a tradeoff between better style alignment and prompt alignment gains. To address this tradeoff, we propose a novel fine-tuning method called Style Tailoring, which jointly fits the content and style distribution and achieves best tradeoff. Evaluation results show our method improves visual quality by 14%, prompt alignment by 16.2% and scene diversity by 15.3%, compared to prompt engineering the base Emu model for stickers generation.

comment: 10 pages, 5 figures

♻ ☆ CTSpine1K: A Large-Scale Dataset for Spinal Vertebrae Segmentation in Computed Tomography MICCAI2024

Yang Deng, Ce Wang, Yuan Hui, Qian Li, Jun Li, Shiwei Luo, Mengke Sun, Quan Quan, Shuxin Yang, You Hao, Pengbo Liu, Honghu Xiao, Chunpeng Zhao, Xinbao Wu, S. Kevin Zhou

Spine-related diseases have high morbidity and cause a huge burden of social cost. Spine imaging is an essential tool for noninvasively visualizing and assessing spinal pathology. Segmenting vertebrae in computed tomography (CT) images is the basis of quantitative medical image analysis for clinical diagnosis and surgery planning of spine diseases. Current publicly available annotated datasets on spinal vertebrae are small in size. Due to the lack of a large-scale annotated spine image dataset, the mainstream deep learning-based segmentation methods, which are data-driven, are heavily restricted. In this paper, we introduce a large-scale spine CT dataset, called CTSpine1K, curated from multiple sources for vertebra segmentation, which contains 1,005 CT volumes with over 11,100 labeled vertebrae belonging to different spinal conditions. Based on this dataset, we conduct several spinal vertebrae segmentation experiments to set the first benchmark. We believe that this large-scale dataset will facilitate further research in many spine-related image analysis tasks, including but not limited to vertebrae segmentation, labeling, 3D spine reconstruction from biplanar radiographs, image super-resolution, and enhancement.

comment: Accepted by MICCAI2024 Open Data for oral presentation and will be published as a part of the journal MELBA special issue

♻ ☆ A Comprehensive Survey of Hallucination in Large Language, Image, Video and Audio Foundation Models EMNLP 2024

Pranab Sahoo, Prabhash Meharia, Akash Ghosh, Sriparna Saha, Vinija Jain, Aman Chadha

comment: EMNLP 2024 Findings

♻ ☆ Identifying and Solving Conditional Image Leakage in Image-to-Video Diffusion Model NeurIPS 2024

Min Zhao, Hongzhou Zhu, Chendong Xiang, Kaiwen Zheng, Chongxuan Li, Jun Zhu

Diffusion models have obtained substantial progress in image-to-video generation. However, in this paper, we find that these models tend to generate videos with less motion than expected. We attribute this to the issue called conditional image leakage, where the image-to-video diffusion models (I2V-DMs) tend to over-rely on the conditional image at large time steps. We further address this challenge from both inference and training aspects. First, we propose to start the generation process from an earlier time step to avoid the unreliable large-time steps of I2V-DMs, as well as an initial noise distribution with optimal analytic expressions (Analytic-Init) by minimizing the KL divergence between it and the actual marginal distribution to bridge the training-inference gap. Second, we design a time-dependent noise distribution (TimeNoise) for the conditional image during training, applying higher noise levels at larger time steps to disrupt it and reduce the model's dependency on it. We validate these general strategies on various I2V-DMs on our collected open-domain image benchmark and the UCF101 dataset. Extensive results show that our methods outperform baselines by producing higher motion scores with lower errors while maintaining image alignment and temporal consistency, thereby yielding superior overall performance and enabling more accurate motion control. The project page: \url{https://cond-image-leak.github.io/}.

comment: NeurIPS 2024. Project page: https://cond-image-leak.github.io/

♻ ☆ Bootstrap3D: Improving Multi-view Diffusion Model with Synthetic Data

Zeyi Sun, Tong Wu, Pan Zhang, Yuhang Zang, Xiaoyi Dong, Yuanjun Xiong, Dahua Lin, Jiaqi Wang

Recent years have witnessed remarkable progress in multi-view diffusion models for 3D content creation. However, there remains a significant gap in image quality and prompt-following ability compared to 2D diffusion models. A critical bottleneck is the scarcity of high-quality 3D objects with detailed captions. To address this challenge, we propose Bootstrap3D, a novel framework that automatically generates an arbitrary quantity of multi-view images to assist in training multi-view diffusion models. Specifically, we introduce a data generation pipeline that employs (1) 2D and video diffusion models to generate multi-view images based on constructed text prompts, and (2) our fine-tuned 3D-aware MV-LLaVA for filtering high-quality data and rewriting inaccurate captions. Leveraging this pipeline, we have generated 1 million high-quality synthetic multi-view images with dense descriptive captions to address the shortage of high-quality 3D data. Furthermore, we present a Training Timestep Reschedule (TTR) strategy that leverages the denoising process to learn multi-view consistency while maintaining the original 2D diffusion prior. Extensive experiments demonstrate that Bootstrap3D can generate high-quality multi-view images with superior aesthetic quality, image-text alignment, and maintained view consistency.

comment: Project Page: https://sunzey.github.io/Bootstrap3D/

♻ ☆ AttackBench: Evaluating Gradient-based Attacks for Adversarial Examples

Antonio Emanuele Cinà, Jérôme Rony, Maura Pintor, Luca Demetrio, Ambra Demontis, Battista Biggio, Ismail Ben Ayed, Fabio Roli

Adversarial examples are typically optimized with gradient-based attacks. While novel attacks are continuously proposed, each is shown to outperform its predecessors using different experimental setups, hyperparameter settings, and number of forward and backward calls to the target models. This provides overly-optimistic and even biased evaluations that may unfairly favor one particular attack over the others. In this work, we aim to overcome these limitations by proposing AttackBench, i.e., the first evaluation framework that enables a fair comparison among different attacks. To this end, we first propose a categorization of gradient-based attacks, identifying their main components and differences. We then introduce our framework, which evaluates their effectiveness and efficiency. We measure these characteristics by (i) defining an optimality metric that quantifies how close an attack is to the optimal solution, and (ii) limiting the number of forward and backward queries to the model, such that all attacks are compared within a given maximum query budget. Our extensive experimental analysis compares more than $100$ attack implementations with a total of over $800$ different configurations against CIFAR-10 and ImageNet models, highlighting that only very few attacks outperform all the competing approaches. Within this analysis, we shed light on several implementation issues that prevent many attacks from finding better solutions or running at all. We release AttackBench as a publicly-available benchmark, aiming to continuously update it to include and evaluate novel gradient-based attacks for optimizing adversarial examples.

comment: https://attackbench.github.io

♻ ☆ I4VGen: Image as Free Stepping Stone for Text-to-Video Generation

Xiefan Guo, Jinlin Liu, Miaomiao Cui, Liefeng Bo, Di Huang

Text-to-video generation has trailed behind text-to-image generation in terms of quality and diversity, primarily due to the inherent complexities of spatio-temporal modeling and the limited availability of video-text datasets. Recent text-to-video diffusion models employ the image as an intermediate step, significantly enhancing overall performance but incurring high training costs. In this paper, we present I4VGen, a novel video diffusion inference pipeline to leverage advanced image techniques to enhance pre-trained text-to-video diffusion models, which requires no additional training. Instead of the vanilla text-to-video inference pipeline, I4VGen consists of two stages: anchor image synthesis and anchor image-augmented text-to-video synthesis. Correspondingly, a simple yet effective generation-selection strategy is employed to achieve visually-realistic and semantically-faithful anchor image, and an innovative noise-invariant video score distillation sampling (NI-VSDS) is developed to animate the image to a dynamic video by distilling motion knowledge from video diffusion models, followed by a video regeneration process to refine the video. Extensive experiments show that the proposed method produces videos with higher visual realism and textual fidelity. Furthermore, I4VGen also supports being seamlessly integrated into existing image-to-video diffusion models, thereby improving overall video quality.

comment: Project page: https://xiefan-guo.github.io/i4vgen

♻ ☆ SSP-RACL: Classification of Noisy Fundus Images with Self-Supervised Pretraining and Robust Adaptive Credal Loss

Mengwen Ye, Yingzi Huangfu, You Li, Zekuan Yu

Fundus image classification is crucial in the computer aided diagnosis tasks, but label noise significantly impairs the performance of deep neural networks. To address this challenge, we propose a robust framework, Self-Supervised Pre-training with Robust Adaptive Credal Loss (SSP-RACL), for handling label noise in fundus image datasets. First, we use Masked Autoencoders (MAE) for pre-training to extract features, unaffected by label noise. Subsequently, RACL employ a superset learning framework, setting confidence thresholds and adaptive label relaxation parameter to construct possibility distributions and provide more reliable ground-truth estimates, thus effectively suppressing the memorization effect. Additionally, we introduce clinical knowledge-based asymmetric noise generation to simulate real-world noisy fundus image datasets. Experimental results demonstrate that our proposed method outperforms existing approaches in handling label noise, showing superior performance.

comment: IEEE BioCAS 2024

♻ ☆ 4K4DGen: Panoramic 4D Generation at 4K Resolution

Renjie Li, Panwang Pan, Bangbang Yang, Dejia Xu, Shijie Zhou, Xuanyang Zhang, Zeming Li, Achuta Kadambi, Zhangyang Wang, Zhengzhong Tu, Zhiwen Fan

The blooming of virtual reality and augmented reality (VR/AR) technologies has driven an increasing demand for the creation of high-quality, immersive, and dynamic environments. However, existing generative techniques either focus solely on dynamic objects or perform outpainting from a single perspective image, failing to meet the requirements of VR/AR applications that need free-viewpoint, 360$^{\circ}$ virtual views where users can move in all directions. In this work, we tackle the challenging task of elevating a single panorama to an immersive 4D experience. For the first time, we demonstrate the capability to generate omnidirectional dynamic scenes with 360$^{\circ}$ views at 4K (4096 $\times$ 2048) resolution, thereby providing an immersive user experience. Our method introduces a pipeline that facilitates natural scene animations and optimizes a set of dynamic Gaussians using efficient splatting techniques for real-time exploration. To overcome the lack of scene-scale annotated 4D data and models, especially in panoramic formats, we propose a novel \textbf{Panoramic Denoiser} that adapts generic 2D diffusion priors to animate consistently in 360$^{\circ}$ images, transforming them into panoramic videos with dynamic scenes at targeted regions. Subsequently, we propose \textbf{Dynamic Panoramic Lifting} to elevate the panoramic video into a 4D immersive environment while preserving spatial and temporal consistency. By transferring prior knowledge from 2D models in the perspective domain to the panoramic domain and the 4D lifting with spatial appearance and geometry regularization, we achieve high-quality Panorama-to-4D generation at a resolution of 4K for the first time.

♻ ☆ LEGO: Learnable Expansion of Graph Operators for Multi-Modal Feature Fusion

Dexuan Ding, Lei Wang, Liyun Zhu, Tom Gedeon, Piotr Koniusz

In computer vision tasks, features often come from diverse representations, domains, and modalities, such as text, images, and videos. Effectively fusing these features is essential for robust performance, especially with the availability of powerful pre-trained models like vision-language models. However, common fusion methods, such as concatenation, element-wise operations, and non-linear techniques, often fail to capture structural relationships, deep feature interactions, and suffer from inefficiency or misalignment of features across domains. In this paper, we shift from high-dimensional feature space to a lower-dimensional, interpretable graph space by constructing similarity graphs that encode feature relationships at different levels, e.g., clip, frame, patch, token, etc. To capture deeper interactions, we use graph power expansions and introduce a learnable graph fusion operator to combine these graph powers for more effective fusion. Our approach is relationship-centric, operates in a homogeneous space, and is mathematically principled, resembling element-wise similarity score aggregation via multilinear polynomials. We demonstrate the effectiveness of our graph-based fusion method on video anomaly detection, showing strong performance across multi-representational, multi-modal, and multi-domain feature fusion tasks.

comment: Research paper

♻ ☆ Pick of the Bunch: Detecting Infrared Small Targets Beyond Hit-Miss Trade-Offs via Selective Rank-Aware Attention

Yimian Dai, Peiwen Pan, Yulei Qian, Yuxuan Li, Xiang Li, Jian Yang, Huan Wang

Infrared small target detection faces the inherent challenge of precisely localizing dim targets amidst complex background clutter. Traditional approaches struggle to balance detection precision and false alarm rates. To break this dilemma, we propose SeRankDet, a deep network that achieves high accuracy beyond the conventional hit-miss trade-off, by following the ``Pick of the Bunch'' principle. At its core lies our Selective Rank-Aware Attention (SeRank) module, employing a non-linear Top-K selection process that preserves the most salient responses, preventing target signal dilution while maintaining constant complexity. Furthermore, we replace the static concatenation typical in U-Net structures with our Large Selective Feature Fusion (LSFF) module, a dynamic fusion strategy that empowers SeRankDet with adaptive feature integration, enhancing its ability to discriminate true targets from false alarms. The network's discernment is further refined by our Dilated Difference Convolution (DDC) module, which merges differential convolution aimed at amplifying subtle target characteristics with dilated convolution to expand the receptive field, thereby substantially improving target-background separation. Despite its lightweight architecture, the proposed SeRankDet sets new benchmarks in state-of-the-art performance across multiple public datasets. The code is available at https://github.com/GrokCV/SeRankDet.

comment: IEEE TGRS 2024

♻ ☆ Representation Synthesis by Probabilistic Many-Valued Logic Operation in Self-Supervised Learning ICIP2024

Hiroki Nakamura, Masashi Okada, Tadahiro Taniguchi

In this paper, we propose a new self-supervised learning (SSL) method for representations that enable logic operations. Representation learning has been applied to various tasks, such as image generation and retrieval. The logical controllability of representations is important for these tasks. Although some methods have been shown to enable the intuitive control of representations using natural languages as the inputs, representation control via logic operations between representations has not been demonstrated. Some SSL methods using representation synthesis (e.g., elementwise mean and maximum operations) have been proposed, but the operations performed in these methods do not incorporate logic operations. In this work, we propose a logic-operable self-supervised representation learning method by replacing the existing representation synthesis with the OR operation on the probabilistic extension of many-valued logic. The representations comprise a set of feature-possession degrees, which are truth values indicating the presence or absence of each feature in the image, and realize the logic operations (e.g., OR and AND). Our method can generate a representation that has the features of both representations or only those features common to both representations. In addition, the expression of the ambiguous presence of a feature is realized by indicating the feature-possession degree by the probability distribution of truth values of the many-valued logic. We showed that our method performs competitively in single and multi-label classification tasks compared with prior SSL methods using synthetic representations. Moreover, experiments on image retrieval using MNIST and PascalVOC showed that the representations of our method can be operated by OR and AND operations.

comment: Accepted to the IEEE Open Journal of Signal Processing (ICIP2024 track)

♻ ☆ Semantic-Aware Adversarial Training for Reliable Deep Hashing Retrieval

Xu Yuan, Zheng Zhang, Xunguang Wang, Lin Wu

Deep hashing has been intensively studied and successfully applied in large-scale image retrieval systems due to its efficiency and effectiveness. Recent studies have recognized that the existence of adversarial examples poses a security threat to deep hashing models, that is, adversarial vulnerability. Notably, it is challenging to efficiently distill reliable semantic representatives for deep hashing to guide adversarial learning, and thereby it hinders the enhancement of adversarial robustness of deep hashing-based retrieval models. Moreover, current researches on adversarial training for deep hashing are hard to be formalized into a unified minimax structure. In this paper, we explore Semantic-Aware Adversarial Training (SAAT) for improving the adversarial robustness of deep hashing models. Specifically, we conceive a discriminative mainstay features learning (DMFL) scheme to construct semantic representatives for guiding adversarial learning in deep hashing. Particularly, our DMFL with the strict theoretical guarantee is adaptively optimized in a discriminative learning manner, where both discriminative and semantic properties are jointly considered. Moreover, adversarial examples are fabricated by maximizing the Hamming distance between the hash codes of adversarial samples and mainstay features, the efficacy of which is validated in the adversarial attack trials. Further, we, for the first time, formulate the formalized adversarial training of deep hashing into a unified minimax optimization under the guidance of the generated mainstay codes. Extensive experiments on benchmark datasets show superb attack performance against the state-of-the-art algorithms, meanwhile, the proposed adversarial training can effectively eliminate adversarial perturbations for trustworthy deep hashing-based retrieval. Our code is available at https://github.com/xandery-geek/SAAT.

♻ ☆ A boundary-aware point clustering approach in Euclidean and embedding spaces for roof plane segmentation

Li Li, Qingqing Li, Guozheng Xu, Pengwei Zhou, Jingmin Tu, Jie Li, Mingming Li, Jian Yao

Roof plane segmentation from airborne LiDAR point clouds is an important technology for 3D building model reconstruction. One of the key issues of plane segmentation is how to design powerful features that can exactly distinguish adjacent planar patches. The quality of point feature directly determines the accuracy of roof plane segmentation. Most of existing approaches use handcrafted features to extract roof planes. However, the abilities of these features are relatively low, especially in boundary area. To solve this problem, we propose a boundary-aware point clustering approach in Euclidean and embedding spaces constructed by a multi-task deep network for roof plane segmentation. We design a three-branch network to predict semantic labels, point offsets and extract deep embedding features. In the first branch, we classify the input data as non-roof, boundary and plane points. In the second branch, we predict point offsets for shifting each point toward its respective instance center. In the third branch, we constrain that points of the same plane instance should have the similar embeddings. We aim to ensure that points of the same plane instance are close as much as possible in both Euclidean and embedding spaces. However, although deep network has strong feature representative ability, it is still hard to accurately distinguish points near plane instance boundary. Therefore, we first group plane points into many clusters in the two spaces, and then we assign the rest boundary points to their closest clusters to generate final complete roof planes. In this way, we can effectively reduce the influence of unreliable boundary points. In addition, we prepare a synthetic dataset and two real datasets to train and evaluate our approach. The experiments results show that the proposed approach significantly outperforms the existing state-of-the-art approaches.

♻ ☆ The Instinctive Bias: Spurious Images lead to Illusion in MLLMs

Tianyang Han, Qing Lian, Rui Pan, Renjie Pi, Jipeng Zhang, Shizhe Diao, Yong Lin, Tong Zhang

Large language models (LLMs) have recently experienced remarkable progress, where the advent of multi-modal large language models (MLLMs) has endowed LLMs with visual capabilities, leading to impressive performances in various multi-modal tasks. However, those powerful MLLMs such as GPT-4V still fail spectacularly when presented with certain image and text inputs. In this paper, we identify a typical class of inputs that baffles MLLMs, which consist of images that are highly relevant but inconsistent with answers, causing MLLMs to suffer from visual illusion. To quantify the effect, we propose CorrelationQA, the first benchmark that assesses the visual illusion level given spurious images. This benchmark contains 7,308 text-image pairs across 13 categories. Based on the proposed CorrelationQA, we conduct a thorough analysis on 9 mainstream MLLMs, illustrating that they universally suffer from this instinctive bias to varying degrees. We hope that our curated benchmark and evaluation results aid in better assessments of the MLLMs' robustness in the presence of misleading images. The code and datasets are available at https://github.com/MasaiahHan/CorrelationQA.

♻ ☆ Rethinking and Defending Protective Perturbation in Personalized Diffusion Models

Yixin Liu, Ruoxi Chen, Xun Chen, Lichao Sun

Personalized diffusion models (PDMs) have become prominent for adapting pretrained text-to-image models to generate images of specific subjects using minimal training data. However, PDMs are susceptible to minor adversarial perturbations, leading to significant degradation when fine-tuned on corrupted datasets. These vulnerabilities are exploited to create protective perturbations that prevent unauthorized image generation. Existing purification methods attempt to mitigate this issue but often over-purify images, resulting in information loss. In this work, we conduct an in-depth analysis of the fine-tuning process of PDMs through the lens of shortcut learning. We hypothesize and empirically demonstrate that adversarial perturbations induce a latent-space misalignment between images and their text prompts in the CLIP embedding space. This misalignment causes the model to erroneously associate noisy patterns with unique identifiers during fine-tuning, resulting in poor generalization. Based on these insights, we propose a systematic defense framework that includes data purification and contrastive decoupling learning. We first employ off-the-shelf image restoration techniques to realign images with their original semantic meanings in latent space. Then, we introduce contrastive decoupling learning with noise tokens to decouple the learning of personalized concepts from spurious noise patterns. Our study not only uncovers fundamental shortcut learning vulnerabilities in PDMs but also provides a comprehensive evaluation framework for developing stronger protection. Our extensive evaluation demonstrates its superiority over existing purification methods and stronger robustness against adaptive perturbation.

comment: Our code is available at https://github.com/liuyixin-louis/DiffShortcut

♻ ☆ LMOD: A Large Multimodal Ophthalmology Dataset and Benchmark for Large Vision-Language Models

Zhenyue Qin, Yu Yin, Dylan Campbell, Xuansheng Wu, Ke Zou, Yih-Chung Tham, Ninghao Liu, Xiuzhen Zhang, Qingyu Chen

Ophthalmology relies heavily on detailed image analysis for diagnosis and treatment planning. While large vision-language models (LVLMs) have shown promise in understanding complex visual information, their performance on ophthalmology images remains underexplored. We introduce LMOD, a dataset and benchmark for evaluating LVLMs on ophthalmology images, covering anatomical understanding, diagnostic analysis, and demographic extraction. LMODincludes 21,993 images spanning optical coherence tomography, scanning laser ophthalmoscopy, eye photos, surgical scenes, and color fundus photographs. We benchmark 13 state-of-the-art LVLMs and find that they are far from perfect for comprehending ophthalmology images. Models struggle with diagnostic analysis and demographic extraction, reveal weaknesses in spatial reasoning, diagnostic analysis, handling out-of-domain queries, and safeguards for handling biomarkers of ophthalmology images.

♻ ☆ Integrating Large Language Models into a Tri-Modal Architecture for Automated Depression Classification

Santosh V. Patapati

Major Depressive Disorder (MDD) is a pervasive mental health condition that affects 300 million people worldwide. This work presents a novel, BiLSTM-based tri-modal model-level fusion architecture for the binary classification of depression from clinical interview recordings. The proposed architecture incorporates Mel Frequency Cepstral Coefficients, Facial Action Units, and uses a two-shot learning based GPT-4 model to process text data. This is the first work to incorporate large language models into a multi-modal architecture for this task. It achieves impressive results on the DAIC-WOZ AVEC 2016 Challenge cross-validation split and Leave-One-Subject-Out cross-validation split, surpassing all baseline models and multiple state-of-the-art models. In Leave-One-Subject-Out testing, it achieves an accuracy of 91.01%, an F1-Score of 85.95%, a precision of 80%, and a recall of 92.86%.

comment: Keywords: Multi-Modal Neural Networks, Deep Learning, Large Language Models, Depression Diagnosis, Biomedical Informatics, DAIC-WOZ

♻ ☆ Towards Data-and Knowledge-Driven Artificial Intelligence: A Survey on Neuro-Symbolic Computing

Wenguan Wang, Yi Yang, Fei Wu

Neural-symbolic computing (NeSy), which pursues the integration of the symbolic and statistical paradigms of cognition, has been an active research area of Artificial Intelligence (AI) for many years. As NeSy shows promise of reconciling the advantages of reasoning and interpretability of symbolic representation and robust learning in neural networks, it may serve as a catalyst for the next generation of AI. In the present paper, we provide a systematic overview of the recent developments and important contributions of NeSy research. Firstly, we introduce study history of this area, covering early work and foundations. We further discuss background concepts and identify key driving factors behind the development of NeSy. Afterward, we categorize recent landmark approaches along several main characteristics that underline this research paradigm, including neural-symbolic integration, knowledge representation, knowledge embedding, and functionality. Next, we briefly discuss the successful application of modern NeSy approaches in several domains. Then, we benchmark several NeSy methods on three representative application tasks. Finally, we identify the open problems together with potential future research directions. This survey is expected to help new researchers enter this rapidly evolving field and accelerate the progress towards data-and knowledge-driven AI.

comment: PAMI 2024

♻ ☆ Color Equivariant Network CVPR 2024

Felix O'Mahony, Yulong Yang, Christine Allen-Blanchette

Group equivariant convolutional neural networks have been designed for a variety of geometric transformations from 2D and 3D rotation groups, to semi-groups such as scale. Despite the improved interpretability, accuracy and generalizability afforded by these architectures, group equivariant networks have seen limited application in the context of perceptual quantities such as hue and saturation, even though their variation can lead to significant reductions in classification performance. In this paper, we introduce convolutional neural networks equivariant to variations in hue and saturation by design. To achieve this, we leverage the observation that hue and saturation transformations can be identified with the 2D rotation and 1D translation groups respectively. Our hue-, saturation-, and fully color-equivariant networks achieve equivariance to these perceptual transformations without an increase in network parameters. We demonstrate the utility of our networks on synthetic and real world datasets where color and lighting variations are commonplace.

comment: Accepted at CVPR 2024 Equivariant Vision Workshop

♻ ☆ Explainable Concept Generation through Vision-Language Preference Learning

Aditya Taparia, Som Sagar, Ransalu Senanayake

Concept-based explanations have become a popular choice for explaining deep neural networks post-hoc because, unlike most other explainable AI techniques, they can be used to test high-level visual "concepts" that are not directly related to feature attributes. For instance, the concept of "stripes" is important to classify an image as a zebra. Concept-based explanation methods, however, require practitioners to guess and collect multiple candidate concept image sets, which can often be imprecise and labor-intensive. Addressing this limitation, in this paper, we frame concept image set creation as an image generation problem. However, since naively using a generative model does not result in meaningful concepts, we devise a reinforcement learning-based preference optimization (RLPO) algorithm that fine-tunes the vision-language generative model from approximate textual descriptions of concepts. Through a series of experiments, we demonstrate the capability of our method to articulate complex and abstract concepts which aligns with the test class that are otherwise challenging to craft manually. In addition to showing the efficacy and reliability of our method, we show how our method can be used as a diagnostic tool for analyzing neural networks.

comment: 25 pages, 27 figures

♻ ☆ EDADepth: Enhanced Data Augmentation for Monocular Depth Estimation

Nischal Khanal, Shivanand Venkanna Sheshappanavar

Due to their text-to-image synthesis feature, diffusion models have recently seen a rise in visual perception tasks, such as depth estimation. The lack of good-quality datasets makes the extraction of a fine-grain semantic context challenging for the diffusion models. The semantic context with fewer details further worsens the process of creating effective text embeddings that will be used as input for diffusion models. In this paper, we propose a novel EDADepth, an enhanced data augmentation method to estimate monocular depth without using additional training data. We use Swin2SR, a super-resolution model, to enhance the quality of input images. We employ the BEiT pre-trained semantic segmentation model for better extraction of text embeddings. We use BLIP-2 tokenizer to generate tokens from these text embeddings. The novelty of our approach is the introduction of Swin2SR, the BEiT model, and the BLIP-2 tokenizer in the diffusion-based pipeline for the monocular depth estimation. Our model achieves state-of-the-art results (SOTA) on the delta3 metric on NYUv2 and KITTI datasets. It also achieves results comparable to those of the SOTA models in the RMSE and REL metrics. Finally, we also show improvements in the visualization of the estimated depth compared to the SOTA diffusion-based monocular depth estimation models. Code: https://github.com/edadepthmde/EDADepth_ICMLA.

♻ ☆ Statistical Test on Diffusion Model-based Anomaly Detection by Selective Inference

Teruyuki Katsuoka, Tomohiro Shiraishi, Daiki Miwa, Vo Nguyen Le Duy, Ichiro Takeuchi

Advancements in AI image generation, particularly diffusion models, have progressed rapidly. However, the absence of an established framework for quantifying the reliability of AI-generated images hinders their use in critical decision-making tasks, such as medical image diagnosis. In this study, we address the task of detecting anomalous regions in medical images using diffusion models and propose a statistical method to quantify the reliability of the detected anomalies. The core concept of our method involves a selective inference framework, wherein statistical tests are conducted under the condition that the images are produced by a diffusion model. With our approach, the statistical significance of anomaly detection results can be quantified in the form of a $p$-value, enabling decision-making with controlled error rates, as is standard in medical practice. We demonstrate the theoretical soundness and practical effectiveness of our statistical test through numerical experiments on both synthetic and brain image datasets.

comment: 30 pages, 7 figures

♻ ☆ PCEvE: Part Contribution Evaluation Based Model Explanation for Human Figure Drawing Assessment and Beyond

Jongseo Lee, Geo Ahn, Seong Tae Kim, Jinwoo Choi

For automatic human figure drawing (HFD) assessment tasks, such as diagnosing autism spectrum disorder (ASD) using HFD images, the clarity and explainability of a model decision are crucial. Existing pixel-level attribution-based explainable AI (XAI) approaches demand considerable effort from users to interpret the semantic information of a region in an image, which can be often time-consuming and impractical. To overcome this challenge, we propose a part contribution evaluation based model explanation (PCEvE) framework. On top of the part detection, we measure the Shapley Value of each individual part to evaluate the contribution to a model decision. Unlike existing attribution-based XAI approaches, the PCEvE provides a straightforward explanation of a model decision, i.e., a part contribution histogram. Furthermore, the PCEvE expands the scope of explanations beyond the conventional sample-level to include class-level and task-level insights, offering a richer, more comprehensive understanding of model behavior. We rigorously validate the PCEvE via extensive experiments on multiple HFD assessment datasets. Also, we sanity-check the proposed method with a set of controlled experiments. Additionally, we demonstrate the versatility and applicability of our method to other domains by applying it to a photo-realistic dataset, the Stanford Cars.

comment: This papaer is under review

♻ ☆ Neural Light Spheres for Implicit Image Stitching and View Synthesis

Ilya Chugunov, Amogh Joshi, Kiran Murthy, Francois Bleibel, Felix Heide

Challenging to capture, and challenging to display on a cellphone screen, the panorama paradoxically remains both a staple and underused feature of modern mobile camera applications. In this work we address both of these challenges with a spherical neural light field model for implicit panoramic image stitching and re-rendering; able to accommodate for depth parallax, view-dependent lighting, and local scene motion and color changes during capture. Fit during test-time to an arbitrary path panoramic video capture -- vertical, horizontal, random-walk -- these neural light spheres jointly estimate the camera path and a high-resolution scene reconstruction to produce novel wide field-of-view projections of the environment. Our single-layer model avoids expensive volumetric sampling, and decomposes the scene into compact view-dependent ray offset and color components, with a total model size of 80 MB per scene, and real-time (50 FPS) rendering at 1080p resolution. We demonstrate improved reconstruction quality over traditional image stitching and radiance field methods, with significantly higher tolerance to scene motion and non-ideal capture settings.

comment: Project site: https://light.princeton.edu/publication/neuls/

♻ ☆ What Do You See? Enhancing Zero-Shot Image Classification with Multimodal Large Language Models

Abdelrahman Abdelhamed, Mahmoud Afifi, Alec Go

Large language models (LLMs) have been effectively used for many computer vision tasks, including image classification. In this paper, we present a simple yet effective approach for zero-shot image classification using multimodal LLMs. By employing multimodal LLMs, we generate comprehensive textual representations from input images. These textual representations are then utilized to generate fixed-dimensional features in a cross-modal embedding space. Subsequently, these features are fused together to perform zero-shot classification using a linear classifier. Our method does not require prompt engineering for each dataset; instead, we use a single, straightforward, set of prompts across all datasets. We evaluated our method on several datasets, and our results demonstrate its remarkable effectiveness, surpassing benchmark accuracy on multiple datasets. On average, our method achieved an accuracy gain of 4.1 percentage points, with an increase of 6.8 percentage points on the ImageNet dataset, compared to prior methods. Our findings highlight the potential of multimodal LLMs to enhance computer vision tasks such as zero-shot image classification, offering a significant improvement over traditional methods.

♻ ☆ OmniBench: Towards The Future of Universal Omni-Language Models

Yizhi Li, Ge Zhang, Yinghao Ma, Ruibin Yuan, Kang Zhu, Hangyu Guo, Yiming Liang, Jiaheng Liu, Zekun Wang, Jian Yang, Siwei Wu, Xingwei Qu, Jinjie Shi, Xinyue Zhang, Zhenzhu Yang, Xiangzhou Wang, Zhaoxiang Zhang, Zachary Liu, Emmanouil Benetos, Wenhao Huang, Chenghua Lin

Recent advancements in multimodal large language models (MLLMs) have aimed to integrate and interpret data across diverse modalities. However, the capacity of these models to concurrently process and reason about multiple modalities remains inadequately explored, partly due to the lack of comprehensive modality-wise benchmarks. We introduce OmniBench, a novel benchmark designed to rigorously evaluate models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define models capable of such tri-modal processing as omni-language models (OLMs). OmniBench is distinguished by high-quality human annotations, ensuring that accurate responses require integrated understanding and reasoning across all three modalities. Our main findings reveal that: i) most OLMs exhibit critical limitations in instruction-following and reasoning capabilities within tri-modal contexts; and ii) most baselines models perform poorly (below 50\% accuracy) even when provided with alternative textual representations of images or/and audio. These results suggest that the ability to construct a consistent context from text, image, and audio is often overlooked in existing MLLM training paradigms. To address this gap, we curate an instruction tuning dataset of 84.5K training samples, OmniInstruct, for training OLMs to adapt to multimodal contexts. We advocate for future research to focus on developing more robust tri-modal integration techniques and training strategies to enhance OLM performance across diverse modalities. The codes and live leaderboard could be found at https://m-a-p.ai/OmniBench.

♻ ☆ ESREAL: Exploiting Semantic Reconstruction to Mitigate Hallucinations in Vision-Language Models ECCV 2024

Minchan Kim, Minyeong Kim, Junik Bae, Suhwan Choi, Sungkyung Kim, Buru Chang

Hallucinations in vision-language models pose a significant challenge to their reliability, particularly in the generation of long captions. Current methods fall short of accurately identifying and mitigating these hallucinations. To address this issue, we introduce ESREAL, a novel unsupervised learning framework designed to suppress the generation of hallucinations through accurate localization and penalization of hallucinated tokens. Initially, ESREAL creates a reconstructed image based on the generated caption and aligns its corresponding regions with those of the original image. This semantic reconstruction aids in identifying both the presence and type of token-level hallucinations within the generated caption. Subsequently, ESREAL computes token-level hallucination scores by assessing the semantic similarity of aligned regions based on the type of hallucination. Finally, ESREAL employs a proximal policy optimization algorithm, where it selectively penalizes hallucinated tokens according to their token-level hallucination scores. Our framework notably reduces hallucinations in LLaVA, InstructBLIP, and mPLUG-Owl2 by 32.81%, 27.08%, and 7.46% on the CHAIR metric. This improvement is achieved solely through signals derived from the image itself, without the need for any image-text pairs.

comment: ECCV 2024

♻ ☆ Controlling the Fidelity and Diversity of Deep Generative Models via Pseudo Density

Shuangqi Li, Chen Liu, Tong Zhang, Hieu Le, Sabine Süsstrunk, Mathieu Salzmann

We introduce an approach to bias deep generative models, such as GANs and diffusion models, towards generating data with either enhanced fidelity or increased diversity. Our approach involves manipulating the distribution of training and generated data through a novel metric for individual samples, named pseudo density, which is based on the nearest-neighbor information from real samples. Our approach offers three distinct techniques to adjust the fidelity and diversity of deep generative models: 1) Per-sample perturbation, enabling precise adjustments for individual samples towards either more common or more unique characteristics; 2) Importance sampling during model inference to enhance either fidelity or diversity in the generated data; 3) Fine-tuning with importance sampling, which guides the generative model to learn an adjusted distribution, thus controlling fidelity and diversity. Furthermore, our fine-tuning method demonstrates the ability to improve the Frechet Inception Distance (FID) for pre-trained generative models with minimal iterations.

♻ ☆ Mora: Enabling Generalist Video Generation via A Multi-Agent Framework

Zhengqing Yuan, Yixin Liu, Yihan Cao, Weixiang Sun, Haolong Jia, Ruoxi Chen, Zhaoxu Li, Bin Lin, Li Yuan, Lifang He, Chi Wang, Yanfang Ye, Lichao Sun

Text-to-video generation has made significant strides, but replicating the capabilities of advanced systems like OpenAI Sora remains challenging due to their closed-source nature. Existing open-source methods struggle to achieve comparable performance, often hindered by ineffective agent collaboration and inadequate training data quality. In this paper, we introduce Mora, a novel multi-agent framework that leverages existing open-source modules to replicate Sora functionalities. We address these fundamental limitations by proposing three key techniques: (1) multi-agent fine-tuning with a self-modulation factor to enhance inter-agent coordination, (2) a data-free training strategy that uses large models to synthesize training data, and (3) a human-in-the-loop mechanism combined with multimodal large language models for data filtering to ensure high-quality training datasets. Our comprehensive experiments on six video generation tasks demonstrate that Mora achieves performance comparable to Sora on VBench, outperforming existing open-source methods across various tasks. Specifically, in the text-to-video generation task, Mora achieved a Video Quality score of 0.800, surpassing Sora 0.797 and outperforming all other baseline models across six key metrics. Additionally, in the image-to-video generation task, Mora achieved a perfect Dynamic Degree score of 1.00, demonstrating exceptional capability in enhancing motion realism and achieving higher Imaging Quality than Sora. These results highlight the potential of collaborative multi-agent systems and human-in-the-loop mechanisms in advancing text-to-video generation. Our code is available at \url{https://github.com/lichao-sun/Mora}.

♻ ☆ CommVQA: Situating Visual Question Answering in Communicative Contexts EMNLP 2024

Nandita Shankar Naik, Christopher Potts, Elisa Kreiss

Current visual question answering (VQA) models tend to be trained and evaluated on image-question pairs in isolation. However, the questions people ask are dependent on their informational needs and prior knowledge about the image content. To evaluate how situating images within naturalistic contexts shapes visual questions, we introduce CommVQA, a VQA dataset consisting of images, image descriptions, real-world communicative scenarios where the image might appear (e.g., a travel website), and follow-up questions and answers conditioned on the scenario and description. CommVQA, which contains 1000 images and 8,949 question-answer pairs, poses a challenge for current models. Error analyses and a human-subjects study suggest that generated answers still contain high rates of hallucinations, fail to fittingly address unanswerable questions, and don't suitably reflect contextual information. Overall, we show that access to contextual information is essential for solving CommVQA, leading to the highest performing VQA model and highlighting the relevance of situating systems within communicative scenarios.

comment: EMNLP 2024 camera ready version

♻ ☆ GaNI: Global and Near Field Illumination Aware Neural Inverse Rendering

Jiaye Wu, Saeed Hadadan, Geng Lin, Matthias Zwicker, David Jacobs, Roni Sengupta

In this paper, we present GaNI, a Global and Near-field Illumination-aware neural inverse rendering technique that can reconstruct geometry, albedo, and roughness parameters from images of a scene captured with co-located light and camera. Existing inverse rendering techniques with co-located light-camera focus on single objects only, without modeling global illumination and near-field lighting more prominent in scenes with multiple objects. We introduce a system that solves this problem in two stages; we first reconstruct the geometry powered by neural volumetric rendering NeuS, followed by inverse neural radiosity that uses the previously predicted geometry to estimate albedo and roughness. However, such a naive combination fails and we propose multiple technical contributions that enable this two-stage approach. We observe that NeuS fails to handle near-field illumination and strong specular reflections from the flashlight in a scene. We propose to implicitly model the effects of near-field illumination and introduce a surface angle loss function to handle specular reflections. Similarly, we observe that invNeRad assumes constant illumination throughout the capture and cannot handle moving flashlights during capture. We propose a light position-aware radiance cache network and additional smoothness priors on roughness to reconstruct reflectance. Experimental evaluation on synthetic and real data shows that our method outperforms the existing co-located light-camera-based inverse rendering techniques. Our approach produces significantly better reflectance and slightly better geometry than capture strategies that do not require a dark room.

♻ ☆ Updating CLIP to Prefer Descriptions Over Captions

Amir Zur, Elisa Kreiss, Karel D'Oosterlinck, Christopher Potts, Atticus Geiger

Although CLIPScore is a powerful generic metric that captures the similarity between a text and an image, it fails to distinguish between a caption that is meant to complement the information in an image and a description that is meant to replace an image entirely, e.g., for accessibility. We address this shortcoming by updating the CLIP model with the Concadia dataset to assign higher scores to descriptions than captions using parameter efficient fine-tuning and a loss objective derived from work on causal interpretability. This model correlates with the judgements of blind and low-vision people while preserving transfer capabilities and has interpretable structure that sheds light on the caption--description distinction.

♻ ☆ MonoVisual3DFilter: 3D tomatoes' localisation with monocular cameras using histogram filters

Sandro Costa Magalhães, Filipe Neves dos Santos, António Paulo Moreira, Jorge Dias

Performing tasks in agriculture, such as fruit monitoring or harvesting, requires perceiving the objects' spatial position. RGB-D cameras are limited under open-field environments due to lightning interferences. So, in this study, we state to answer the research question: "How can we use and control monocular sensors to perceive objects' position in the 3D task space?" Towards this aim, we approached histogram filters (Bayesian discrete filters) to estimate the position of tomatoes in the tomato plant through the algorithm MonoVisual3DFilter. Two kernel filters were studied: the square kernel and the Gaussian kernel. The implemented algorithm was essayed in simulation, with and without Gaussian noise and random noise, and in a testbed at laboratory conditions. The algorithm reported a mean absolute error lower than 10 mm in simulation and 20 mm in the testbed at laboratory conditions with an assessing distance of about 0.5 m. So, the results are viable for real environments and should be improved at closer distances.

♻ ☆ ERIC: Estimating Rainfall with Commodity Doorbell Camera for Precision Residential Irrigation

Tian Liu, Liuyi Jin, Radu Stoleru, Amran Haroon, Charles Swanson, Kexin Feng

Current state-of-the-art residential irrigation systems, such as WaterMyYard, rely on rainfall data from nearby weather stations to adjust irrigation amounts. However, the accuracy of rainfall data is compromised by the limited spatial resolution of rain gauges and the significant variability of hyperlocal rainfall, leading to substantial water waste. To improve irrigation efficiency, we developed a cost-effective irrigation system, dubbed ERIC, which employs machine learning models to estimate rainfall from commodity doorbell camera footage and optimizes irrigation schedules without human intervention. Specifically, we: a) designed novel visual and audio features with lightweight neural network models to infer rainfall from the camera at the edge, preserving user privacy; b) built a complete end-to-end irrigation system on Raspberry Pi 4, costing only \$75. We deployed the system across five locations (collecting over 750 hours of video) with varying backgrounds and light conditions. Comprehensive evaluation validates that ERIC achieves state-of-the-art rainfall estimation performance ($\sim$ 5mm/day), saving 9,112 gallons/month of water, translating to \$28.56/month in utility savings. Data and code are available at https://github.com/LENSS/ERIC-BuildSys2024.git

comment: BuildSys 2024

♻ ☆ Visually Descriptive Language Model for Vector Graphics Reasoning

Zhenhailong Wang, Joy Hsu, Xingyao Wang, Kuan-Hao Huang, Manling Li, Jiajun Wu, Heng Ji

Despite significant advancements, large multimodal models (LMMs) still struggle to bridge the gap between low-level visual perception -- focusing on shapes, sizes, and layouts -- and high-level language reasoning, such as semantics and logic. This limitation is evident in tasks that require precise visual perception, like comparing geometric properties or solving visual reasoning problems. To study this failure mode, we focus on vector graphics -- images composed of 2D objects and shapes, prevalent in LMM-based tasks in web, design, and OS environments. We identify two key research questions: how can we enable precise visual perception, and how can we facilitate high-level reasoning based on such low-level perceptions? To capture fine visual details, we use Scalable Vector Graphics (SVG) for accurate encoding of visual scenes. However, SVGs are not readily interpretable by LMMs in a zero-shot manner. To tackle this, we propose the Visually Descriptive Language Model (VDLM), which introduces a Primal Visual Description (PVD) as an intermediate textual representation. PVD translates SVGs into a text-based abstraction consisting of primitive attributes (e.g., shape, position, measurement) and their corresponding values. PVD can be learned using task-agnostic synthesized data and represents visual primitives that are universal across vector graphics. This abstraction is more structured, allowing for direct interpretation by foundation models for zero-shot generalization. Without human-annotated data, empirical results show that VDLM significantly improves state-of-the-art LMMs like GPT-4o on various multimodal perception and reasoning tasks. Extensive analyses of VDLM show improved interpretability due to its disentangled perception and reasoning. We also demonstrate a positive correlation between PVD quality and task performance. Project page: https://mikewangwzhl.github.io/VDLM/

comment: Project page: https://mikewangwzhl.github.io/VDLM/

♻ ☆ VALOR-EVAL: Holistic Coverage and Faithfulness Evaluation of Large Vision-Language Models ACL 2024

Haoyi Qiu, Wenbo Hu, Zi-Yi Dou, Nanyun Peng

Large Vision-Language Models (LVLMs) suffer from hallucination issues, wherein the models generate plausible-sounding but factually incorrect outputs, undermining their reliability. A comprehensive quantitative evaluation is necessary to identify and understand the extent of hallucinations in these models. However, existing benchmarks are often limited in scope, focusing mainly on object hallucinations. Furthermore, current evaluation methods struggle to effectively address the subtle semantic distinctions between model outputs and reference data, as well as the balance between hallucination and informativeness. To address these issues, we introduce a multi-dimensional benchmark covering objects, attributes, and relations, with challenging images selected based on associative biases. Moreover, we propose a large language model (LLM)-based two-stage evaluation framework that generalizes the popular CHAIR metric and incorporates both faithfulness and coverage into the evaluation. Experiments on 10 established LVLMs demonstrate that our evaluation metric is more comprehensive and better correlated with humans than existing work when evaluating on our challenging human-annotated benchmark dataset. Our work also highlights the critical balance between faithfulness and coverage of model outputs, and encourages future works to address hallucinations in LVLMs while keeping their outputs informative.

comment: ACL 2024 Findings

♻ ☆ MosaicFusion: Diffusion Models as Data Augmenters for Large Vocabulary Instance Segmentation

Jiahao Xie, Wei Li, Xiangtai Li, Ziwei Liu, Yew Soon Ong, Chen Change Loy

We present MosaicFusion, a simple yet effective diffusion-based data augmentation approach for large vocabulary instance segmentation. Our method is training-free and does not rely on any label supervision. Two key designs enable us to employ an off-the-shelf text-to-image diffusion model as a useful dataset generator for object instances and mask annotations. First, we divide an image canvas into several regions and perform a single round of diffusion process to generate multiple instances simultaneously, conditioning on different text prompts. Second, we obtain corresponding instance masks by aggregating cross-attention maps associated with object prompts across layers and diffusion time steps, followed by simple thresholding and edge-aware refinement processing. Without bells and whistles, our MosaicFusion can produce a significant amount of synthetic labeled data for both rare and novel categories. Experimental results on the challenging LVIS long-tailed and open-vocabulary benchmarks demonstrate that MosaicFusion can significantly improve the performance of existing instance segmentation models, especially for rare and novel categories. Code: https://github.com/Jiahao000/MosaicFusion.

comment: International Journal of Computer Vision (IJCV), 2024

♻ ☆ GroundingBooth: Grounding Text-to-Image Customization

Zhexiao Xiong, Wei Xiong, Jing Shi, He Zhang, Yizhi Song, Nathan Jacobs

Recent studies in text-to-image customization show great success in generating personalized object variants given several images of a subject. While existing methods focus more on preserving the identity of the subject, they often fall short of controlling the spatial relationship between objects. In this work, we introduce GroundingBooth, a framework that achieves zero-shot instance-level spatial grounding on both foreground subjects and background objects in the text-to-image customization task. Our proposed text-image grounding module and masked cross-attention layer allow us to generate personalized images with both accurate layout alignment and identity preservation while maintaining text-image coherence. With such layout control, our model inherently enables the customization of multiple subjects at once. Our model is evaluated on both layout-guided image synthesis and reference-based customization tasks, showing strong results compared to existing methods. Our work is the first work to achieve a joint grounding on both subject-driven foreground generation and text-driven background generation.

♻ ☆ WaveMixSR-V2: Enhancing Super-resolution with Higher Efficiency

Pranav Jeevan, Neeraj Nixon, Amit Sethi

Recent advancements in single image super-resolution have been predominantly driven by token mixers and transformer architectures. WaveMixSR utilized the WaveMix architecture, employing a two-dimensional discrete wavelet transform for spatial token mixing, achieving superior performance in super-resolution tasks with remarkable resource efficiency. In this work, we present an enhanced version of the WaveMixSR architecture by (1) replacing the traditional transpose convolution layer with a pixel shuffle operation and (2) implementing a multistage design for higher resolution tasks ($4\times$). Our experiments demonstrate that our enhanced model -- WaveMixSR-V2 -- outperforms other architectures in multiple super-resolution tasks, achieving state-of-the-art for the BSD100 dataset, while also consuming fewer resources, exhibits higher parameter efficiency, lower latency and higher throughput. Our code is available at https://github.com/pranavphoenix/WaveMixSR.

comment: 10 pages. arXiv admin note: text overlap with arXiv:2307.00430

♻ ☆ AURA : Automatic Mask Generator using Randomized Input Sampling for Object Removal

Changsuk Oh, H. Jin Kim

The objective of the image inpainting task is to fill missing regions of an image in a visually plausible way. Recently, deep-learning-based image inpainting networks have generated outstanding results, and some utilize their models as object removers by masking unwanted objects in an image. However, while trying to better remove objects using their networks, the previous works pay less attention to the importance of the input mask. In this paper, we focus on generating the input mask to better remove objects using the off-the-shelf image inpainting network. We propose an automatic mask generator inspired by the explainable AI (XAI) method, whose output can better remove objects than a semantic segmentation mask. The proposed method generates an importance map using randomly sampled input masks and quantitatively estimated scores of the completed images obtained from the random masks. The output mask is selected by a judge module among the candidate masks which are generated from the importance map. We design the judge module to quantitatively estimate the quality of the object removal results. In addition, we empirically find that the evaluation methods used in the previous works reporting object removal results are not appropriate for estimating the performance of an object remover. Therefore, we propose new evaluation metrics (FID$^*$ and U-IDS$^*$) to properly evaluate the quality of object removers. Experiments confirm that our method shows better performance in removing target class objects than the masks generated from the semantic segmentation maps, and the two proposed metrics make judgments consistent with humans.

♻ ☆ EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis

Alexander Mai, Peter Hedman, George Kopanas, Dor Verbin, David Futschik, Qiangeng Xu, Falko Kuester, Jonathan T. Barron, Yinda Zhang

We present Exact Volumetric Ellipsoid Rendering (EVER), a method for real-time differentiable emission-only volume rendering. Unlike recent rasterization based approach by 3D Gaussian Splatting (3DGS), our primitive based representation allows for exact volume rendering, rather than alpha compositing 3D Gaussian billboards. As such, unlike 3DGS our formulation does not suffer from popping artifacts and view dependent density, but still achieves frame rates of $\sim\!30$ FPS at 720p on an NVIDIA RTX4090. Since our approach is built upon ray tracing it enables effects such as defocus blur and camera distortion (e.g. such as from fisheye cameras), which are difficult to achieve by rasterization. We show that our method is more accurate with fewer blending issues than 3DGS and follow-up work on view-consistent rendering, especially on the challenging large-scale scenes from the Zip-NeRF dataset where it achieves sharpest results among real-time techniques.

comment: Project page: https://half-potato.gitlab.io/posts/ever

♻ ☆ Mitigating Dialogue Hallucination for Large Vision Language Models via Adversarial Instruction Tuning

Dongmin Park, Zhaofang Qian, Guangxing Han, Ser-Nam Lim

Mitigating hallucinations of Large Vision Language Models,(LVLMs) is crucial to enhance their reliability for general-purpose assistants. This paper shows that such hallucinations of LVLMs can be significantly exacerbated by preceding user-system dialogues. To precisely measure this, we first present an evaluation benchmark by extending popular multi-modal benchmark datasets with prepended hallucinatory dialogues powered by our novel Adversarial Question Generator (AQG), which can automatically generate image-related yet adversarial dialogues by adopting adversarial attacks on LVLMs. On our benchmark, the zero-shot performance of state-of-the-art LVLMs drops significantly for both the VQA and Captioning tasks. Next, we further reveal this hallucination is mainly due to the prediction bias toward preceding dialogues rather than visual content. To reduce this bias, we propose Adversarial Instruction Tuning (AIT) that robustly fine-tunes LVLMs against hallucinatory dialogues. Extensive experiments show our proposed approach successfully reduces dialogue hallucination while maintaining performance.

♻ ☆ Not (yet) the whole story: Evaluating Visual Storytelling Requires More than Measuring Coherence, Grounding, and Repetition EMNLP 2024

Aditya K Surikuchi, Raquel Fernández, Sandro Pezzelle

Visual storytelling consists in generating a natural language story given a temporally ordered sequence of images. This task is not only challenging for models, but also very difficult to evaluate with automatic metrics since there is no consensus about what makes a story 'good'. In this paper, we introduce a novel method that measures story quality in terms of human likeness regarding three key aspects highlighted in previous work: visual grounding, coherence, and repetitiveness. We then use this method to evaluate the stories generated by several models, showing that the foundation model LLaVA obtains the best result, but only slightly so compared to TAPM, a 50-times smaller visual storytelling model. Upgrading the visual and language components of TAPM results in a model that yields competitive performance with a relatively low number of parameters. Finally, we carry out a human evaluation study, whose results suggest that a 'good' story may require more than a human-like level of visual grounding, coherence, and repetition.

comment: In proceedings of EMNLP 2024 (Findings)

Machine Learning 150

☆ Flash-Splat: 3D Reflection Removal with Flash Cues and Gaussian Splats

Mingyang Xie, Haoming Cai, Sachin Shah, Yiran Xu, Brandon Y. Feng, Jia-Bin Huang, Christopher A. Metzler

☆ Vinoground: Scrutinizing LMMs over Dense Temporal Reasoning with Short Videos

Jianrui Zhang, Mu Cai, Yong Jae Lee

comment: Project Page: https://vinoground.github.io

☆ Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations

Nick Jiang, Anish Kachinthaya, Suzie Petryk, Yossi Gandelsman

comment: Project page and code: http://anishk23733.github.io/vl-interp/

☆ Erasing Conceptual Knowledge from Language Models

Rohit Gandikota, Sheridan Feucht, Samuel Marks, David Bau

comment: Project Page: https://elm.baulab.info

☆ Forecasting Smog Clouds With Deep Learning

Valentijn Oldenburg, Juan Cardenas-Cartagena, Matias Valdenegro-Toro

In this proof-of-concept study, we conduct multivariate timeseries forecasting for the concentrations of nitrogen dioxide (NO2), ozone (O3), and (fine) particulate matter (PM10 & PM2.5) with meteorological covariates between two locations using various deep learning models, with a focus on long short-term memory (LSTM) and gated recurrent unit (GRU) architectures. In particular, we propose an integrated, hierarchical model architecture inspired by air pollution dynamics and atmospheric science that employs multi-task learning and is benchmarked by unidirectional and fully-connected models. Results demonstrate that, above all, the hierarchical GRU proves itself as a competitive and efficient method for forecasting the concentration of smog-related pollutants.

☆ SIEVE: General Purpose Data Filtering System Matching GPT-4o Accuracy at 1% the Cost

Jifan Zhang, Robert Nowak

☆ ReLIC: A Recipe for 64k Steps of In-Context Reinforcement Learning for Embodied AI

Ahmad Elawady, Gunjan Chhablani, Ram Ramrakhya, Karmesh Yadav, Dhruv Batra, Zsolt Kira, Andrew Szot

Intelligent embodied agents need to quickly adapt to new scenarios by integrating long histories of experience into decision-making. For instance, a robot in an unfamiliar house initially wouldn't know the locations of objects needed for tasks and might perform inefficiently. However, as it gathers more experience, it should learn the layout of its environment and remember where objects are, allowing it to complete new tasks more efficiently. To enable such rapid adaptation to new tasks, we present ReLIC, a new approach for in-context reinforcement learning (RL) for embodied agents. With ReLIC, agents are capable of adapting to new environments using 64,000 steps of in-context experience with full attention while being trained through self-generated experience via RL. We achieve this by proposing a novel policy update scheme for on-policy RL called "partial updates'' as well as a Sink-KV mechanism that enables effective utilization of a long observation history for embodied agents. Our method outperforms a variety of meta-RL baselines in adapting to unseen houses in an embodied multi-object navigation task. In addition, we find that ReLIC is capable of few-shot imitation learning despite never being trained with expert demonstrations. We also provide a comprehensive analysis of ReLIC, highlighting that the combination of large-scale RL training, the proposed partial updates scheme, and the Sink-KV are essential for effective in-context learning. The code for ReLIC and all our experiments is at https://github.com/aielawady/relic

☆ An Online Automatic Modulation Classification Scheme Based on Isolation Distributional Kernel

Xinpeng Li, Zile Jiang, Kai Ming Ting, Ye Zhu

Automatic Modulation Classification (AMC), as a crucial technique in modern non-cooperative communication networks, plays a key role in various civil and military applications. However, existing AMC methods usually are complicated and can work in batch mode only due to their high computational complexity. This paper introduces a new online AMC scheme based on Isolation Distributional Kernel. Our method stands out in two aspects. Firstly, it is the first proposal to represent baseband signals using a distributional kernel. Secondly, it introduces a pioneering AMC technique that works well in online settings under realistic time-varying channel conditions. Through extensive experiments in online settings, we demonstrate the effectiveness of the proposed classifier. Our results indicate that the proposed approach outperforms existing baseline models, including two state-of-the-art deep learning classifiers. Moreover, it distinguishes itself as the first online classifier for AMC with linear time complexity, which marks a significant efficiency boost for real-time applications.

☆ Training Language Models on Synthetic Edit Sequences Improves Code Synthesis

Ulyana Piterbarg, Lerrel Pinto, Rob Fergus

☆ CriSPO: Multi-Aspect Critique-Suggestion-guided Automatic Prompt Optimization for Text Generation

Han He, Qianchu Liu, Lei Xu, Chaitanya Shivade, Yi Zhang, Sundararajan Srinivasan, Katrin Kirchhoff

☆ Contrastive Localized Language-Image Pre-Training

Hong-You Chen, Zhengfeng Lai, Haotian Zhang, Xinze Wang, Marcin Eichner, Keen You, Meng Cao, Bowen Zhang, Yinfei Yang, Zhe Gan

comment: Preprint

☆ Neutral residues: revisiting adapters for model extension

Franck Signe Talla, Herve Jegou, Edouard Grave

☆ Grounding Large Language Models In Embodied Environment With Imperfect World Models

Haolan Liu, Jishen Zhao

☆ Salient Information Prompting to Steer Content in Prompt-based Abstractive Summarization EMNLP 2024

Lei Xu, Mohammed Asad Karim, Saket Dingliwal, Aparna Elangovan

Large language models (LLMs) can generate fluent summaries across domains using prompting techniques, reducing the need to train models for summarization applications. However, crafting effective prompts that guide LLMs to generate summaries with the appropriate level of detail and writing style remains a challenge. In this paper, we explore the use of salient information extracted from the source document to enhance summarization prompts. We show that adding keyphrases in prompts can improve ROUGE F1 and recall, making the generated summaries more similar to the reference and more complete. The number of keyphrases can control the precision-recall trade-off. Furthermore, our analysis reveals that incorporating phrase-level salient information is superior to word- or sentence-level. However, the impact on hallucination is not universally positive across LLMs. To conduct this analysis, we introduce Keyphrase Signal Extractor (SigExt), a lightweight model that can be finetuned to extract salient keyphrases. By using SigExt, we achieve consistent ROUGE improvements across datasets and open-weight and proprietary LLMs without any LLM customization. Our findings provide insights into leveraging salient information in building prompt-based summarization systems.

comment: Accepted to EMNLP 2024 Industry Track

☆ Revisit Large-Scale Image-Caption Data in Pre-training Multimodal Foundation Models

Zhengfeng Lai, Vasileios Saveris, Chen Chen, Hong-You Chen, Haotian Zhang, Bowen Zhang, Juan Lao Tebar, Wenze Hu, Zhe Gan, Peter Grasch, Meng Cao, Yinfei Yang

comment: CV/ML

☆ OOD-Chameleon: Is Algorithm Selection for OOD Generalization Learnable?

Liangze Jiang, Damien Teney

Out-of-distribution (OOD) generalization is challenging because distribution shifts come in many forms. A multitude of learning algorithms exist and each can improve performance in specific OOD situations. We posit that much of the challenge of OOD generalization lies in choosing the right algorithm for the right dataset. However, such algorithm selection is often elusive under complex real-world shifts. In this work, we formalize the task of algorithm selection for OOD generalization and investigate whether it could be approached by learning. We propose a solution, dubbed OOD-Chameleon that treats the task as a supervised classification over candidate algorithms. We construct a dataset of datasets to learn from, which represents diverse types, magnitudes and combinations of shifts (covariate shift, label shift, spurious correlations). We train the model to predict the relative performance of algorithms given a dataset's characteristics. This enables a priori selection of the best learning strategy, i.e. without training various models as needed with traditional model selection. Our experiments show that the adaptive selection outperforms any individual algorithm and simple selection heuristics, on unseen datasets of controllable and realistic image data. Inspecting the model shows that it learns non-trivial data/algorithms interactions, and reveals the conditions for any one algorithm to surpass another. This opens new avenues for (1) enhancing OOD generalization with existing algorithms instead of designing new ones, and (2) gaining insights into the applicability of existing algorithms with respect to datasets' properties.

☆ Data Similarity-Based One-Shot Clustering for Multi-Task Hierarchical Federated Learning

Abdulmoneam Ali, Ahmed Arafa

We address the problem of cluster identity estimation in a hierarchical federated learning setting in which users work toward learning different tasks. To overcome the challenge of task heterogeneity, users need to be grouped in a way such that users with the same task are in the same group, conducting training together, while sharing the weights of feature extraction layers with the other groups. Toward that end, we propose a one-shot clustering algorithm that can effectively identify and group users based on their data similarity. This enables more efficient collaboration and sharing of a common layer representation within the federated learning system. Our proposed algorithm not only enhances the clustering process, but also overcomes challenges related to privacy concerns, communication overhead, and the need for prior knowledge about learning models or loss function behaviors. We validate our proposed algorithm using various datasets such as CIFAR-10 and Fashion MNIST, and show that it outperforms the baseline in terms of accuracy and variance reduction.

comment: To appear in Asilomar 2024

☆ Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation

Rohin Manvi, Anikait Singh, Stefano Ermon

☆ Large Language Models as Markov Chains

Oussama Zekri, Ambroise Odonnat, Abdelhakim Benechehab, Linus Bleistein, Nicolas Boullé, Ievgen Redko

comment: 49 pages, 17 figures

☆ SynthFormer: Equivariant Pharmacophore-based Generation of Molecules for Ligand-Based Drug Design

Zygimantas Jocys, Henriette M. G. Willems, Katayoun Farrahi

Drug discovery is a complex and resource-intensive process, with significant time and cost investments required to bring new medicines to patients. Recent advancements in generative machine learning (ML) methods offer promising avenues to accelerate early-stage drug discovery by efficiently exploring chemical space. This paper addresses the gap between in silico generative approaches and practical in vitro methodologies, highlighting the need for their integration to optimize molecule discovery. We introduce SynthFormer, a novel ML model that utilizes a 3D equivariant encoder for pharmacophores to generate fully synthesizable molecules, constructed as synthetic trees. Unlike previous methods, SynthFormer incorporates 3D information and provides synthetic paths, enhancing its ability to produce molecules with good docking scores across various proteins. Our contributions include a new methodology for efficient chemical space exploration using 3D information, a novel architecture called Synthformer for translating 3D pharmacophore representations into molecules, and a meaningful embedding space that organizes reagents for drug discovery optimization. Synthformer generates molecules that dock well and enables effective late-stage optimization restricted by synthesis paths.

☆ Measurements with Noise: Bayesian Optimization for Co-optimizing Noise and Property Discovery in Automated Experiments

Boris N. Slautin, Yu Liu, Jan Dec, Vladimir V. Shvartsman, Doru C. Lupascu, Maxim Ziatdinov, Sergei V. Kalinin

We have developed a Bayesian optimization (BO) workflow that integrates intra-step noise optimization into automated experimental cycles. Traditional BO approaches in automated experiments focus on optimizing experimental trajectories but often overlook the impact of measurement noise on data quality and cost. Our proposed framework simultaneously optimizes both the target property and the associated measurement noise by introducing time as an additional input parameter, thereby balancing the signal-to-noise ratio and experimental duration. Two approaches are explored: a reward-driven noise optimization and a double-optimization acquisition function, both enhancing the efficiency of automated workflows by considering noise and cost within the optimization process. We validate our method through simulations and real-world experiments using Piezoresponse Force Microscopy (PFM), demonstrating the successful optimization of measurement duration and property exploration. Our approach offers a scalable solution for optimizing multiple variables in automated experimental workflows, improving data quality, and reducing resource expenditure in materials science and beyond.

comment: 22 pages, 9 figures

☆ AlzhiNet: Traversing from 2DCNN to 3DCNN, Towards Early Detection and Diagnosis of Alzheimer's Disease

Romoke Grace Akindele, Samuel Adebayo, Paul Shekonya Kanda, Ming Yu

☆ NETS: A Non-Equilibrium Transport Sampler

Michael S. Albergo, Eric Vanden-Eijnden

We propose an algorithm, termed the Non-Equilibrium Transport Sampler (NETS), to sample from unnormalized probability distributions. NETS can be viewed as a variant of annealed importance sampling (AIS) based on Jarzynski's equality, in which the stochastic differential equation used to perform the non-equilibrium sampling is augmented with an additional learned drift term that lowers the impact of the unbiasing weights used in AIS. We show that this drift is the minimizer of a variety of objective functions, which can all be estimated in an unbiased fashion without backpropagating through solutions of the stochastic differential equations governing the sampling. We also prove that some these objectives control the Kullback-Leibler divergence of the estimated distribution from its target. NETS is shown to be unbiased and, in addition, has a tunable diffusion coefficient which can be adjusted post-training to maximize the effective sample size. We demonstrate the efficacy of the method on standard benchmarks, high-dimensional Gaussian mixture distributions, and a model from statistical lattice field theory, for which it surpasses the performances of related work and existing baselines.

☆ Selective Attention Improves Transformer

Yaniv Leviathan, Matan Kalman, Yossi Matias

☆ Lie Algebra Canonicalization: Equivariant Neural Operators under arbitrary Lie Groups

Zakhar Shumaylov, Peter Zaika, James Rowbottom, Ferdia Sherry, Melanie Weber, Carola-Bibiane Schönlieb

comment: 40 pages; preprint

☆ Discovering Clues of Spoofed LM Watermarks

Thibaud Gloaguen, Nikola Jovanović, Robin Staab, Martin Vechev

LLM watermarks stand out as a promising way to attribute ownership of LLM-generated text. One threat to watermark credibility comes from spoofing attacks, where an unauthorized third party forges the watermark, enabling it to falsely attribute arbitrary texts to a particular LLM. While recent works have demonstrated that state-of-the-art schemes are in fact vulnerable to spoofing, they lack deeper qualitative analysis of the texts produced by spoofing methods. In this work, we for the first time reveal that there are observable differences between genuine and spoofed watermark texts. Namely, we show that regardless of their underlying approach, all current spoofing methods consistently leave observable artifacts in spoofed texts, indicative of watermark forgery. We build upon these findings to propose rigorous statistical tests that reliably reveal the presence of such artifacts, effectively discovering that a watermark was spoofed. Our experimental evaluation shows high test power across all current spoofing methods, providing insights into their fundamental limitations, and suggesting a way to mitigate this threat.

☆ DailyDilemmas: Revealing Value Preferences of LLMs with Quandaries of Daily Life

Yu Ying Chiu, Liwei Jiang, Yejin Choi

comment: Preprint. Under Review

☆ Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Models

Shuoyuan Wang, Yixuan Li, Hongxin Wei

Confidence calibration is critical for the safe deployment of machine learning models in the real world. However, such issue in vision-language models like CLIP, particularly after fine-tuning, has not been fully addressed. In this work, we demonstrate that existing prompt tuning methods usually lead to a trade-off of calibration between base and new classes: the cross-entropy loss in CoOp causes overconfidence in new classes by increasing textual label divergence, whereas the regularization of KgCoOp maintains the confidence level but results in underconfidence in base classes due to the improved accuracy. Inspired by the observations, we introduce Dynamic Outlier Regularization (DOR) to ensure the confidence calibration on both base and new classes after fine-tuning. In particular, we propose to minimize the feature deviation of novel textual labels (instead of base classes) sampled from a large vocabulary. In effect, DOR prevents the increase in textual divergence for new labels while easing restrictions on base classes. Extensive experiments demonstrate that DOR can enhance the calibration performance of current fine-tuning methods on base and new classes.

comment: Preprint

☆ Highly Adaptive Ridge

Alejandro Schuler, Alexander Hagemeister, Mark van der Laan

In this paper we propose the Highly Adaptive Ridge (HAR): a regression method that achieves a $n^{-1/3}$ dimension-free L2 convergence rate in the class of right-continuous functions with square-integrable sectional derivatives. This is a large nonparametric function class that is particularly appropriate for tabular data. HAR is exactly kernel ridge regression with a specific data-adaptive kernel based on a saturated zero-order tensor-product spline basis expansion. We use simulation and real data to confirm our theory. We demonstrate empirical performance better than state-of-the-art algorithms for small datasets in particular.

☆ CulturalBench: a Robust, Diverse and Challenging Benchmark on Measuring the (Lack of) Cultural Knowledge of LLMs

Yu Ying Chiu, Liwei Jiang, Bill Yuchen Lin, Chan Young Park, Shuyue Stella Li, Sahithya Ravi, Mehar Bhatia, Maria Antoniak, Yulia Tsvetkov, Vered Shwartz, Yejin Choi

comment: Preprint. Under review

☆ FAN: Fourier Analysis Networks

Yihong Dong, Ge Li, Yongding Tao, Xue Jiang, Kechi Zhang, Jia Li, Jing Su, Jun Zhang, Jingjing Xu

☆ GUD: Generation with Unified Diffusion

Mathis Gerdes, Max Welling, Miranda C. N. Cheng

Diffusion generative models transform noise into data by inverting a process that progressively adds noise to data samples. Inspired by concepts from the renormalization group in physics, which analyzes systems across different scales, we revisit diffusion models by exploring three key design aspects: 1) the choice of representation in which the diffusion process operates (e.g. pixel-, PCA-, Fourier-, or wavelet-basis), 2) the prior distribution that data is transformed into during diffusion (e.g. Gaussian with covariance $\Sigma$), and 3) the scheduling of noise levels applied separately to different parts of the data, captured by a component-wise noise schedule. Incorporating the flexibility in these choices, we develop a unified framework for diffusion generative models with greatly enhanced design freedom. In particular, we introduce soft-conditioning models that smoothly interpolate between standard diffusion models and autoregressive models (in any basis), conceptually bridging these two approaches. Our framework opens up a wide design space which may lead to more efficient training and data generation, and paves the way to novel architectures integrating different generative approaches and generation tasks.

comment: 11 pages, 8 figures

☆ AlphaIntegrator: Transformer Action Search for Symbolic Integration Proofs

Mert Ünsal, Timon Gehr, Martin Vechev

We present the first correct-by-construction learning-based system for step-by-step mathematical integration. The key idea is to learn a policy, represented by a GPT transformer model, which guides the search for the right mathematical integration rule, to be carried out by a symbolic solver. Concretely, we introduce a symbolic engine with axiomatically correct actions on mathematical expressions, as well as the first dataset for step-by-step integration. Our GPT-style transformer model, trained on this synthetic data, demonstrates strong generalization by surpassing its own data generator in accuracy and efficiency, using 50% fewer search steps. Our experimental results with SoTA LLMs also demonstrate that the standard approach of fine-tuning LLMs on a set of question-answer pairs is insufficient for solving this mathematical task. This motivates the importance of discovering creative methods for combining LLMs with symbolic reasoning engines, of which our work is an instance.

☆ How to Train Long-Context Language Models (Effectively)

Tianyu Gao, Alexander Wettig, Howard Yen, Danqi Chen

comment: Our code, data, and models are available at https://github.com/princeton-nlp/ProLong

☆ Scalable Simulation-free Entropic Unbalanced Optimal Transport

Jaemoo Choi, Jaewoong Choi

The Optimal Transport (OT) problem investigates a transport map that connects two distributions while minimizing a given cost function. Finding such a transport map has diverse applications in machine learning, such as generative modeling and image-to-image translation. In this paper, we introduce a scalable and simulation-free approach for solving the Entropic Unbalanced Optimal Transport (EUOT) problem. We derive the dynamical form of this EUOT problem, which is a generalization of the Schr\"odinger bridges (SB) problem. Based on this, we derive dual formulation and optimality conditions of the EUOT problem from the stochastic optimal control interpretation. By leveraging these properties, we propose a simulation-free algorithm to solve EUOT, called Simulation-free EUOT (SF-EUOT). While existing SB models require expensive simulation costs during training and evaluation, our model achieves simulation-free training and one-step generation by utilizing the reciprocal property. Our model demonstrates significantly improved scalability in generative modeling and image-to-image translation tasks compared to previous SB methods.

comment: 26 pages

☆ Deconstructing Recurrence, Attention, and Gating: Investigating the transferability of Transformers and Gated Recurrent Neural Networks in forecasting of dynamical systems

Hunter S. Heidenreich, Pantelis R. Vlachas, Petros Koumoutsakos

Machine learning architectures, including transformers and recurrent neural networks (RNNs) have revolutionized forecasting in applications ranging from text processing to extreme weather. Notably, advanced network architectures, tuned for applications such as natural language processing, are transferable to other tasks such as spatiotemporal forecasting tasks. However, there is a scarcity of ablation studies to illustrate the key components that enable this forecasting accuracy. The absence of such studies, although explainable due to the associated computational cost, intensifies the belief that these models ought to be considered as black boxes. In this work, we decompose the key architectural components of the most powerful neural architectures, namely gating and recurrence in RNNs, and attention mechanisms in transformers. Then, we synthesize and build novel hybrid architectures from the standard blocks, performing ablation studies to identify which mechanisms are effective for each task. The importance of considering these components as hyper-parameters that can augment the standard architectures is exhibited on various forecasting datasets, from the spatiotemporal chaotic dynamics of the multiscale Lorenz 96 system, the Kuramoto-Sivashinsky equation, as well as standard real world time-series benchmarks. A key finding is that neural gating and attention improves the performance of all standard RNNs in most tasks, while the addition of a notion of recurrence in transformers is detrimental. Furthermore, our study reveals that a novel, sparsely used, architecture which integrates Recurrent Highway Networks with neural gating and attention mechanisms, emerges as the best performing architecture in high-dimensional spatiotemporal forecasting of dynamical systems.

☆ CAX: Cellular Automata Accelerated in JAX

Maxence Faldor, Antoine Cully

Cellular automata have become a cornerstone for investigating emergence and self-organization across diverse scientific disciplines, spanning neuroscience, artificial life, and theoretical physics. However, the absence of a hardware-accelerated cellular automata library limits the exploration of new research directions, hinders collaboration, and impedes reproducibility. In this work, we introduce CAX (Cellular Automata Accelerated in JAX), a high-performance and flexible open-source library designed to accelerate cellular automata research. CAX offers cutting-edge performance and a modular design through a user-friendly interface, and can support both discrete and continuous cellular automata with any number of dimensions. We demonstrate CAX's performance and flexibility through a wide range of benchmarks and applications. From classic models like elementary cellular automata and Conway's Game of Life to advanced applications such as growing neural cellular automata and self-classifying MNIST digits, CAX speeds up simulations up to 2,000 times faster. Furthermore, we demonstrate CAX's potential to accelerate research by presenting a collection of three novel cellular automata experiments, each implemented in just a few lines of code thanks to the library's modular architecture. Notably, we show that a simple one-dimensional cellular automaton can outperform GPT-4 on the 1D-ARC challenge.

☆ Immunogenicity Prediction with Dual Attention Enables Vaccine Target Selection

Song Li, Yang Tan, Song Ke, Liang Hong, Bingxin Zhou

comment: 18 pages, 11 tables, 5 figures

☆ Labor Migration Modeling through Large-scale Job Query Data

Zhuoning Guo, Le Zhang, Hengshu Zhu, Weijia Zhang, Hui Xiong, Hao Liu

Accurate and timely modeling of labor migration is crucial for various urban governance and commercial tasks, such as local policy-making and business site selection. However, existing studies on labor migration largely rely on limited survey data with statistical methods, which fail to deliver timely and fine-grained insights for time-varying regional trends. To this end, we propose a deep learning-based spatial-temporal labor migration analysis framework, DHG-SIL, by leveraging large-scale job query data. Specifically, we first acquire labor migration intention as a proxy of labor migration via job queries from one of the world's largest search engines. Then, a Disprepant Homophily co-preserved Graph Convolutional Network (DH-GCN) and an interpretable temporal module are respectively proposed to capture cross-city and sequential labor migration dependencies. Besides, we introduce four interpretable variables to quantify city migration properties, which are co-optimized with city representations via tailor-designed contrastive losses. Extensive experiments on three real-world datasets demonstrate the superiority of our DHG-SIL. Notably, DHG-SIL has been deployed as a core component of a cooperative partner's intelligent human resource system, and the system supported a series of city talent attraction reports.

☆ Estimating Generalization Performance Along the Trajectory of Proximal SGD in Robust Regression

Kai Tan, Pierre C. Bellec

This paper studies the generalization performance of iterates obtained by Gradient Descent (GD), Stochastic Gradient Descent (SGD) and their proximal variants in high-dimensional robust regression problems. The number of features is comparable to the sample size and errors may be heavy-tailed. We introduce estimators that precisely track the generalization error of the iterates along the trajectory of the iterative algorithm. These estimators are provably consistent under suitable conditions. The results are illustrated through several examples, including Huber regression, pseudo-Huber regression, and their penalized variants with non-smooth regularizer. We provide explicit generalization error estimates for iterates generated from GD and SGD, or from proximal SGD in the presence of a non-smooth regularizer. The proposed risk estimates serve as effective proxies for the actual generalization error, allowing us to determine the optimal stopping iteration that minimizes the generalization error. Extensive simulations confirm the effectiveness of the proposed generalization error estimates.

☆ Inverse Entropic Optimal Transport Solves Semi-supervised Learning via Data Likelihood Maximization

Mikhail Persiianov, Arip Asadulaev, Nikita Andreev, Nikita Starodubcev, Dmitry Baranchuk, Anastasis Kratsios, Evgeny Burnaev, Alexander Korotin

Learning conditional distributions $\pi^*(\cdot|x)$ is a central problem in machine learning, which is typically approached via supervised methods with paired data $(x,y) \sim \pi^*$. However, acquiring paired data samples is often challenging, especially in problems such as domain translation. This necessitates the development of $\textit{semi-supervised}$ models that utilize both limited paired data and additional unpaired i.i.d. samples $x \sim \pi^*_x$ and $y \sim \pi^*_y$ from the marginal distributions. The usage of such combined data is complex and often relies on heuristic approaches. To tackle this issue, we propose a new learning paradigm that integrates both paired and unpaired data $\textbf{seamlessly}$ through the data likelihood maximization techniques. We demonstrate that our approach also connects intriguingly with inverse entropic optimal transport (OT). This finding allows us to apply recent advances in computational OT to establish a $\textbf{light}$ learning algorithm to get $\pi^*(\cdot|x)$. Furthermore, we demonstrate through empirical tests that our method effectively learns conditional distributions using paired and unpaired data simultaneously.

☆ Online Learning Guided Quasi-Newton Methods with Global Non-Asymptotic Convergence

Ruichen Jiang, Aryan Mokhtari

In this paper, we propose a quasi-Newton method for solving smooth and monotone nonlinear equations, including unconstrained minimization and minimax optimization as special cases. For the strongly monotone setting, we establish two global convergence bounds: (i) a linear convergence rate that matches the rate of the celebrated extragradient method, and (ii) an explicit global superlinear convergence rate that provably surpasses the linear convergence rate after at most ${O}(d)$ iterations, where $d$ is the problem's dimension. In addition, for the case where the operator is only monotone, we prove a global convergence rate of ${O}(\min\{{1}/{k},{\sqrt{d}}/{k^{1.25}}\})$ in terms of the duality gap. This matches the rate of the extragradient method when $k = {O}(d^2)$ and is faster when $k = \Omega(d^2)$. These results are the first global convergence results to demonstrate a provable advantage of a quasi-Newton method over the extragradient method, without querying the Jacobian of the operator. Unlike classical quasi-Newton methods, we achieve this by using the hybrid proximal extragradient framework and a novel online learning approach for updating the Jacobian approximation matrices. Specifically, guided by the convergence analysis, we formulate the Jacobian approximation update as an online convex optimization problem over non-symmetric matrices, relating the regret of the online problem to the convergence rate of our method. To facilitate efficient implementation, we further develop a tailored online learning algorithm based on an approximate separation oracle, which preserves structures such as symmetry and sparsity in the Jacobian matrices.

comment: 54 pages

☆ Diss-l-ECT: Dissecting Graph Data with local Euler Characteristic Transforms

Julius von Rohrscheidt, Bastian Rieck

The Euler Characteristic Transform (ECT) is an efficiently-computable geometrical-topological invariant that characterizes the global shape of data. In this paper, we introduce the Local Euler Characteristic Transform ($\ell$-ECT), a novel extension of the ECT particularly designed to enhance expressivity and interpretability in graph representation learning. Unlike traditional Graph Neural Networks (GNNs), which may lose critical local details through aggregation, the $\ell$-ECT provides a lossless representation of local neighborhoods. This approach addresses key limitations in GNNs by preserving nuanced local structures while maintaining global interpretability. Moreover, we construct a rotation-invariant metric based on $\ell$-ECTs for spatial alignment of data spaces. Our method exhibits superior performance than standard GNNs on a variety of node classification tasks, particularly in graphs with high heterophily.

☆ Achieving Fairness in Predictive Process Analytics via Adversarial Learning (Extended Version)

Massimiliano de Leoni, Alessandro Padella

Predictive business process analytics has become important for organizations, offering real-time operational support for their processes. However, these algorithms often perform unfair predictions because they are based on biased variables (e.g., gender or nationality), namely variables embodying discrimination. This paper addresses the challenge of integrating a debiasing phase into predictive business process analytics to ensure that predictions are not influenced by biased variables. Our framework leverages on adversial debiasing is evaluated on four case studies, showing a significant reduction in the contribution of biased variables to the predicted value. The proposed technique is also compared with the state of the art in fairness in process mining, illustrating that our framework allows for a more enhanced level of fairness, while retaining a better prediction quality.

comment: 17 pages, 5 figures

☆ LoGra-Med: Long Context Multi-Graph Alignment for Medical Vision-Language Model

Duy M. H. Nguyen, Nghiem T. Diep, Trung Q. Nguyen, Hoang-Bao Le, Tai Nguyen, Tien Nguyen, TrungTin Nguyen, Nhat Ho, Pengtao Xie, Roger Wattenhofer, James Zhou, Daniel Sonntag, Mathias Niepert

State-of-the-art medical multi-modal large language models (med-MLLM), like LLaVA-Med or BioMedGPT, leverage instruction-following data in pre-training. However, those models primarily focus on scaling the model size and data volume to boost performance while mainly relying on the autoregressive learning objectives. Surprisingly, we reveal that such learning schemes might result in a weak alignment between vision and language modalities, making these models highly reliant on extensive pre-training datasets - a significant challenge in medical domains due to the expensive and time-consuming nature of curating high-quality instruction-following instances. We address this with LoGra-Med, a new multi-graph alignment algorithm that enforces triplet correlations across image modalities, conversation-based descriptions, and extended captions. This helps the model capture contextual meaning, handle linguistic variability, and build cross-modal associations between visuals and text. To scale our approach, we designed an efficient end-to-end learning scheme using black-box gradient estimation, enabling faster LLaMa 7B training. Our results show LoGra-Med matches LLAVA-Med performance on 600K image-text pairs for Medical VQA and significantly outperforms it when trained on 10% of the data. For example, on VQA-RAD, we exceed LLAVA-Med by 20.13% and nearly match the 100% pre-training score (72.52% vs. 72.64%). We also surpass SOTA methods like BiomedGPT on visual chatbots and RadFM on zero-shot image classification with VQA, highlighting the effectiveness of multi-graph alignment.

comment: First version

☆ IndicSentEval: How Effectively do Multilingual Transformer Models encode Linguistic Properties for Indic Languages?

Akhilesh Aravapalli, Mounika Marreddy, Subba Reddy Oota, Radhika Mamidi, Manish Gupta

comment: 23 pages, 11 figures

☆ Beyond Expected Returns: A Policy Gradient Algorithm for Cumulative Prospect Theoretic Reinforcement Learning

Olivier Lepel, Anas Barakat

The widely used expected utility theory has been shown to be empirically inconsistent with human preferences in the psychology and behavioral economy literatures. Cumulative Prospect Theory (CPT) has been developed to fill in this gap and provide a better model for human-based decision-making supported by empirical evidence. It allows to express a wide range of attitudes and perceptions towards risk, gains and losses. A few years ago, CPT has been combined with Reinforcement Learning (RL) to formulate a CPT policy optimization problem where the goal of the agent is to search for a policy generating long-term returns which are aligned with their preferences. In this work, we revisit this policy optimization problem and provide new insights on optimal policies and their nature depending on the utility function under consideration. We further derive a novel policy gradient theorem for the CPT policy optimization objective generalizing the seminal corresponding result in standard RL. This result enables us to design a model-free policy gradient algorithm to solve the CPT-RL problem. We illustrate the performance of our algorithm in simple examples motivated by traffic control and electricity management applications. We also demonstrate that our policy gradient algorithm scales better to larger state spaces compared to the existing zeroth order algorithm for solving the same problem.

comment: 33 pages, 19 figures

☆ Long-Sequence Recommendation Models Need Decoupled Embeddings

Ningya Feng, Junwei Pan, Jialong Wu, Baixu Chen, Ximei Wang, Qian Li, Xian Hu, Jie Jiang, Mingsheng Long

Lifelong user behavior sequences, comprising up to tens of thousands of history behaviors, are crucial for capturing user interests and predicting user responses in modern recommendation systems. A two-stage paradigm is typically adopted to handle these long sequences: a few relevant behaviors are first searched from the original long sequences via an attention mechanism in the first stage and then aggregated with the target item to construct a discriminative representation for prediction in the second stage. In this work, we identify and characterize, for the first time, a neglected deficiency in existing long-sequence recommendation models: a single set of embeddings struggles with learning both attention and representation, leading to interference between these two processes. Initial attempts to address this issue using linear projections -- a technique borrowed from language processing -- proved ineffective, shedding light on the unique challenges of recommendation models. To overcome this, we propose the Decoupled Attention and Representation Embeddings (DARE) model, where two distinct embedding tables are initialized and learned separately to fully decouple attention and representation. Extensive experiments and analysis demonstrate that DARE provides more accurate search of correlated behaviors and outperforms baselines with AUC gains up to 0.9% on public datasets and notable online system improvements. Furthermore, decoupling embedding spaces allows us to reduce the attention embedding dimension and accelerate the search procedure by 50% without significant performance impact, enabling more efficient, high-performance online serving.

comment: First three authors contributed equally

☆ Agents' Room: Narrative Generation through Multi-step Collaboration ICLR 2025

Fantine Huot, Reinald Kim Amplayo, Jennimaria Palomaki, Alice Shoshana Jakobovits, Elizabeth Clark, Mirella Lapata

comment: Under review as a conference paper at ICLR 2025

☆ Diffusion & Adversarial Schrödinger Bridges via Iterative Proportional Markovian Fitting

Sergei Kholkin, Grigoriy Ksenofontov, David Li, Nikita Kornilov, Nikita Gushchin, Evgeny Burnaev, Alexander Korotin

The Iterative Markovian Fitting (IMF) procedure based on iterative reciprocal and Markovian projections has recently been proposed as a powerful method for solving the Schr\"odinger Bridge problem. However, it has been observed that for the practical implementation of this procedure, it is crucial to alternate between fitting a forward and backward time diffusion at each iteration. Such implementation is thought to be a practical heuristic, which is required to stabilize training and obtain good results in applications such as unpaired domain translation. In our work, we show that this heuristic closely connects with the pioneer approaches for the Schr\"odinger Bridge based on the Iterative Proportional Fitting (IPF) procedure. Namely, we find that the practical implementation of IMF is, in fact, a combination of IMF and IPF procedures, and we call this combination the Iterative Proportional Markovian Fitting (IPMF) procedure. We show both theoretically and practically that this combined IPMF procedure can converge under more general settings, thus, showing that the IPMF procedure opens a door towards developing a unified framework for solving Schr\"odinger Bridge problems.

☆ Three-in-One: Fast and Accurate Transducer for Hybrid-Autoregressive ASR

Hainan Xu, Travis M. Bartley, Vladimir Bataev, Boris Ginsburg

We present \textbf{H}ybrid-\textbf{A}utoregressive \textbf{IN}ference Tr\textbf{AN}sducers (HAINAN), a novel architecture for speech recognition that extends the Token-and-Duration Transducer (TDT) model. Trained with randomly masked predictor network outputs, HAINAN supports both autoregressive inference with all network components and non-autoregressive inference without the predictor. Additionally, we propose a novel semi-autoregressive inference paradigm that first generates an initial hypothesis using non-autoregressive inference, followed by refinement steps where each token prediction is regenerated using parallelized autoregression on the initial hypothesis. Experiments on multiple datasets across different languages demonstrate that HAINAN achieves efficiency parity with CTC in non-autoregressive mode and with TDT in autoregressive mode. In terms of accuracy, autoregressive HAINAN outperforms TDT and RNN-T, while non-autoregressive HAINAN significantly outperforms CTC. Semi-autoregressive inference further enhances the model's accuracy with minimal computational overhead, and even outperforms TDT results in some cases. These results highlight HAINAN's flexibility in balancing accuracy and speed, positioning it as a strong candidate for real-world speech recognition applications.

☆ Beyond Squared Error: Exploring Loss Design for Enhanced Training of Generative Flow Networks

Rui Hu, Yifan Zhang, Zhuoran Li, Longbo Huang

Generative Flow Networks (GFlowNets) are a novel class of generative models designed to sample from unnormalized distributions and have found applications in various important tasks, attracting great research interest in their training algorithms. In general, GFlowNets are trained by fitting the forward flow to the backward flow on sampled training objects. Prior work focused on the choice of training objects, parameterizations, sampling and resampling strategies, and backward policies, aiming to enhance credit assignment, exploration, or exploitation of the training process. However, the choice of regression loss, which can highly influence the exploration and exploitation behavior of the under-training policy, has been overlooked. Due to the lack of theoretical understanding for choosing an appropriate regression loss, most existing algorithms train the flow network by minimizing the squared error of the forward and backward flows in log-space, i.e., using the quadratic regression loss. In this work, we rigorously prove that distinct regression losses correspond to specific divergence measures, enabling us to design and analyze regression losses according to the desired properties of the corresponding divergence measures. Specifically, we examine two key properties: zero-forcing and zero-avoiding, where the former promotes exploitation and higher rewards, and the latter encourages exploration and enhances diversity. Based on our theoretical framework, we propose three novel regression losses, namely, Shifted-Cosh, Linex(1/2), and Linex(1). We evaluate them across three benchmarks: hyper-grid, bit-sequence generation, and molecule generation. Our proposed losses are compatible with most existing training algorithms, and significantly improve the performances of the algorithms concerning convergence speed, sample diversity, and robustness.

☆ IC3M: In-Car Multimodal Multi-object Monitoring for Abnormal Status of Both Driver and Passengers

Zihan Fang, Zheng Lin, Senkang Hu, Hangcheng Cao, Yiqin Deng, Xianhao Chen, Yuguang Fang

comment: 16 pages, 17 figures

☆ Generalization emerges from local optimization in a self-organized learning network

S. Barland, L. Gil

We design and analyze a new paradigm for building supervised learning networks, driven only by local optimization rules without relying on a global error function. Traditional neural networks with a fixed topology are made up of identical nodes and derive their expressiveness from an appropriate adjustment of connection weights. In contrast, our network stores new knowledge in the nodes accurately and instantaneously, in the form of a lookup table. Only then is some of this information structured and incorporated into the network geometry. The training error is initially zero by construction and remains so throughout the network topology transformation phase. The latter involves a small number of local topological transformations, such as splitting or merging of nodes and adding binary connections between them. The choice of operations to be carried out is only driven by optimization of expressivity at the local scale. What we are primarily looking for in a learning network is its ability to generalize, i.e. its capacity to correctly answer questions for which it has never learned the answers. We show on numerous examples of classification tasks that the networks generated by our algorithm systematically reach such a state of perfect generalization when the number of learned examples becomes sufficiently large. We report on the dynamics of the change of state and show that it is abrupt and has the distinctive characteristics of a first order phase transition, a phenomenon already observed for traditional learning networks and known as grokking. In addition to proposing a non-potential approach for the construction of learning networks, our algorithm makes it possible to rethink the grokking transition in a new light, under which acquisition of training data and topological structuring of data are completely decoupled phenomena.

comment: This paper is submitted to Phys. Rev. X. It's a physicist's study that focus on a new paradigm for deep learning networks. We would have liked to choose other keywords for arXiv to reach a wider community, but don't have the rights to do so

☆ Boosting Sample Efficiency and Generalization in Multi-agent Reinforcement Learning via Equivariance NeurIPS 2024

Joshua McClellan, Naveed Haghani, John Winder, Furong Huang, Pratap Tokekar

Multi-Agent Reinforcement Learning (MARL) struggles with sample inefficiency and poor generalization [1]. These challenges are partially due to a lack of structure or inductive bias in the neural networks typically used in learning the policy. One such form of structure that is commonly observed in multi-agent scenarios is symmetry. The field of Geometric Deep Learning has developed Equivariant Graph Neural Networks (EGNN) that are equivariant (or symmetric) to rotations, translations, and reflections of nodes. Incorporating equivariance has been shown to improve learning efficiency and decrease error [ 2 ]. In this paper, we demonstrate that EGNNs improve the sample efficiency and generalization in MARL. However, we also show that a naive application of EGNNs to MARL results in poor early exploration due to a bias in the EGNN structure. To mitigate this bias, we present Exploration-enhanced Equivariant Graph Neural Networks or E2GN2. We compare E2GN2 to other common function approximators using common MARL benchmarks MPE and SMACv2. E2GN2 demonstrates a significant improvement in sample efficiency, greater final reward convergence, and a 2x-5x gain in over standard GNNs in our generalization tests. These results pave the way for more reliable and effective solutions in complex multi-agent systems.

comment: accepted as a poster at NeurIPS 2024

☆ Deep Learning-Based Prediction of Suspension Dynamics Performance in Multi-Axle Vehicles

Kai Chun Lin, Bo-Yi Lin

This paper presents a deep learning-based framework for predicting the dynamic performance of suspension systems in multi-axle vehicles, emphasizing the integration of machine learning with traditional vehicle dynamics modeling. A Multi-Task Deep Belief Network Deep Neural Network (MTL-DBN-DNN) was developed to capture the relationships between key vehicle parameters and suspension performance metrics. The model was trained on data generated from numerical simulations and demonstrated superior prediction accuracy compared to conventional DNN models. A comprehensive sensitivity analysis was conducted to assess the impact of various vehicle and suspension parameters on dynamic suspension performance. Additionally, the Suspension Dynamic Performance Index (SDPI) was introduced as a holistic measure to quantify overall suspension performance, accounting for the combined effects of multiple parameters. The findings highlight the effectiveness of multitask learning in improving predictive models for complex vehicle systems.

☆ The Benefit of Being Bayesian in Online Conformal Prediction

Zhiyu Zhang, Zhou Lu, Heng Yang

Based on the framework of Conformal Prediction (CP), we study the online construction of valid confidence sets given a black-box machine learning model. By converting the target confidence levels into quantile levels, the problem can be reduced to predicting the quantiles (in hindsight) of a sequentially revealed data sequence. Two very different approaches have been studied previously. (i) Direct approach: Assuming the data sequence is iid or exchangeable, one could maintain the empirical distribution of the observed data as an algorithmic belief, and directly predict its quantiles. (ii) Indirect approach: As statistical assumptions often do not hold in practice, a recent trend is to consider the adversarial setting and apply first-order online optimization to moving quantile losses (Gibbs & Cand\`es, 2021). It requires knowing the target quantile level beforehand, and suffers from certain validity issues on the obtained confidence sets, due to the associated loss linearization. This paper presents a novel Bayesian CP framework that combines their strengths. Without any statistical assumption, it is able to both: (i) answer multiple arbitrary confidence level queries online, with provably low regret; and (ii) overcome the validity issues suffered by first-order optimization baselines, due to being "data-centric" rather than "iterate-centric". From a technical perspective, our key idea is to regularize the algorithmic belief of the above direct approach by a Bayesian prior, which "robustifies" it by simulating a non-linearized Follow the Regularized Leader (FTRL) algorithm on the output. For statisticians, this can be regarded as an online adversarial view of Bayesian inference. Importantly, the proposed belief update backbone is shared by prediction heads targeting different confidence levels, bringing practical benefits analogous to U-calibration (Kleinberg et al., 2023).

☆ Obtaining Lower Query Complexities through Lightweight Zeroth-Order Proximal Gradient Algorithms

Bin Gu, Xiyuan Wei, Hualin Zhang, Yi Chang, Heng Huang

Zeroth-order (ZO) optimization is one key technique for machine learning problems where gradient calculation is expensive or impossible. Several variance reduced ZO proximal algorithms have been proposed to speed up ZO optimization for non-smooth problems, and all of them opted for the coordinated ZO estimator against the random ZO estimator when approximating the true gradient, since the former is more accurate. While the random ZO estimator introduces bigger error and makes convergence analysis more challenging compared to coordinated ZO estimator, it requires only $\mathcal{O}(1)$ computation, which is significantly less than $\mathcal{O}(d)$ computation of the coordinated ZO estimator, with $d$ being dimension of the problem space. To take advantage of the computationally efficient nature of the random ZO estimator, we first propose a ZO objective decrease (ZOOD) property which can incorporate two different types of errors in the upper bound of convergence rate. Next, we propose two generic reduction frameworks for ZO optimization which can automatically derive the convergence results for convex and non-convex problems respectively, as long as the convergence rate for the inner solver satisfies the ZOOD property. With the application of two reduction frameworks on our proposed ZOR-ProxSVRG and ZOR-ProxSAGA, two variance reduced ZO proximal algorithms with fully random ZO estimators, we improve the state-of-the-art function query complexities from $\mathcal{O}\left(\min\{\frac{dn^{1/2}}{\epsilon^2}, \frac{d}{\epsilon^3}\}\right)$ to $\tilde{\mathcal{O}}\left(\frac{n+d}{\epsilon^2}\right)$ under $d > n^{\frac{1}{2}}$ for non-convex problems, and from $\mathcal{O}\left(\frac{d}{\epsilon^2}\right)$ to $\tilde{\mathcal{O}}\left(n\log\frac{1}{\epsilon}+\frac{d}{\epsilon}\right)$ for convex problems.

comment: Neural Computation 36 (5), 897-935

☆ ColaCare: Enhancing Electronic Health Record Modeling through Large Language Model-Driven Multi-Agent Collaboration

Zixiang Wang, Yinghao Zhu, Huiya Zhao, Xiaochen Zheng, Tianlong Wang, Wen Tang, Yasha Wang, Chengwei Pan, Ewen M. Harrison, Junyi Gao, Liantao Ma

☆ Local Flow Matching Generative Models

Chen Xu, Xiuyuan Cheng, Yao Xie

Flow Matching (FM) is a simulation-free method for learning a continuous and invertible flow to interpolate between two distributions, and in particular to generate data from noise in generative modeling. In this paper, we introduce Local Flow Matching (LFM), which learns a sequence of FM sub-models and each matches a diffusion process up to the time of the step size in the data-to-noise direction. In each step, the two distributions to be interpolated by the sub-model are closer to each other than data vs. noise, and this enables the use of smaller models with faster training. The stepwise structure of LFM is natural to be distilled and different distillation techniques can be adopted to speed up generation. Theoretically, we prove a generation guarantee of the proposed flow model in terms of the $\chi^2$-divergence between the generated and true data distributions. In experiments, we demonstrate the improved training efficiency and competitive generative performance of LFM compared to FM on the unconditional generation of tabular data and image datasets, and also on the conditional generation of robotic manipulation policies.

☆ Diffusion Models are Evolutionary Algorithms

Yanbo Zhang, Benedikt Hartl, Hananel Hazan, Michael Levin

☆ Fair Decentralized Learning

Sayan Biswas, Anne-Marie Kermarrec, Rishi Sharma, Thibaud Trinca, Martijn de Vos

☆ Semantic-Guided RL for Interpretable Feature Engineering

Mohamed Bouadi, Arta Alavi, Salima Benbernou, Mourad Ouziri

The quality of Machine Learning (ML) models strongly depends on the input data, as such generating high-quality features is often required to improve the predictive accuracy. This process is referred to as Feature Engineering (FE). However, since manual feature engineering is time-consuming and requires case-by-case domain knowledge, Automated Feature Engineering (AutoFE) is crucial. A major challenge that remains is to generate interpretable features. To tackle this problem, we introduce SMART, a hybrid approach that uses semantic technologies to guide the generation of interpretable features through a two-step process: Exploitation and Exploration. The former uses Description Logics (DL) to reason on the semantics embedded in Knowledge Graphs (KG) to infer domain-specific features, while the latter exploits the knowledge graph to conduct a guided exploration of the search space through Deep Reinforcement Learning (DRL). Our experiments on public datasets demonstrate that SMART significantly improves prediction accuracy while ensuring a high level of interpretability.

comment: arXiv admin note: substantial text overlap with arXiv:2406.00544

☆ Learning Emergence of Interaction Patterns across Independent RL Agents in Multi-Agent Environments

Vasanth Reddy Baddam, Suat Gumussoy, Almuatazbellah Boker, Hoda Eldardiry

Many real-world problems, such as controlling swarms of drones and urban traffic, naturally lend themselves to modeling as multi-agent reinforcement learning (RL) problems. However, existing multi-agent RL methods often suffer from scalability challenges, primarily due to the introduction of communication among agents. Consequently, a key challenge lies in adapting the success of deep learning in single-agent RL to the multi-agent setting. In response to this challenge, we propose an approach that fundamentally reimagines multi-agent environments. Unlike conventional methods that model each agent individually with separate networks, our approach, the Bottom Up Network (BUN), adopts a unique perspective. BUN treats the collective of multi-agents as a unified entity while employing a specialized weight initialization strategy that promotes independent learning. Furthermore, we dynamically establish connections among agents using gradient information, enabling coordination when necessary while maintaining these connections as limited and sparse to effectively manage the computational budget. Our extensive empirical evaluations across a variety of cooperative multi-agent scenarios, including tasks such as cooperative navigation and traffic control, consistently demonstrate BUN's superiority over baseline methods with substantially reduced computational costs.

comment: 13 pages, 24 figures

☆ Minimax Group Fairness in Strategic Classification

Emily Diana, Saeed Sharifi-Malvajerdi, Ali Vakilian

In strategic classification, agents manipulate their features, at a cost, to receive a positive classification outcome from the learner's classifier. The goal of the learner in such settings is to learn a classifier that is robust to strategic manipulations. While the majority of works in this domain consider accuracy as the primary objective of the learner, in this work, we consider learning objectives that have group fairness guarantees in addition to accuracy guarantees. We work with the minimax group fairness notion that asks for minimizing the maximal group error rate across population groups. We formalize a fairness-aware Stackelberg game between a population of agents consisting of several groups, with each group having its own cost function, and a learner in the agnostic PAC setting in which the learner is working with a hypothesis class H. When the cost functions of the agents are separable, we show the existence of an efficient algorithm that finds an approximately optimal deterministic classifier for the learner when the number of groups is small. This algorithm remains efficient, both statistically and computationally, even when H is the set of all classifiers. We then consider cost functions that are not necessarily separable and show the existence of oracle-efficient algorithms that find approximately optimal randomized classifiers for the learner when H has finite strategic VC dimension. These algorithms work under the assumption that the learner is fully transparent: the learner draws a classifier from its distribution (randomized classifier) before the agents respond by manipulating their feature vectors. We highlight the effectiveness of such transparency in developing oracle-efficient algorithms. We conclude with verifying the efficacy of our algorithms on real data by conducting an experimental analysis.

☆ SAFLEX: Self-Adaptive Augmentation via Feature Label Extrapolation ICLR 2024

Mucong Ding, Bang An, Yuancheng Xu, Anirudh Satheesh, Furong Huang

Data augmentation, a cornerstone technique in deep learning, is crucial in enhancing model performance, especially with scarce labeled data. While traditional techniques are effective, their reliance on hand-crafted methods limits their applicability across diverse data types and tasks. Although modern learnable augmentation methods offer increased adaptability, they are computationally expensive and challenging to incorporate within prevalent augmentation workflows. In this work, we present a novel, efficient method for data augmentation, effectively bridging the gap between existing augmentation strategies and emerging datasets and learning tasks. We introduce SAFLEX (Self-Adaptive Augmentation via Feature Label EXtrapolation), which learns the sample weights and soft labels of augmented samples provided by any given upstream augmentation pipeline, using a specifically designed efficient bilevel optimization algorithm. Remarkably, SAFLEX effectively reduces the noise and label errors of the upstream augmentation pipeline with a marginal computational cost. As a versatile module, SAFLEX excels across diverse datasets, including natural and medical images and tabular data, showcasing its prowess in few-shot learning and out-of-distribution generalization. SAFLEX seamlessly integrates with common augmentation strategies like RandAug, CutMix, and those from large pre-trained generative models like stable diffusion and is also compatible with frameworks such as CLIP's fine-tuning. Our findings highlight the potential to adapt existing augmentation pipelines for new data types and tasks, signaling a move towards more adaptable and resilient training frameworks.

comment: ICLR 2024

☆ Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systems

Guibin Zhang, Yanwei Yue, Zhixun Li, Sukwon Yun, Guancheng Wan, Kun Wang, Dawei Cheng, Jeffrey Xu Yu, Tianlong Chen

Recent advancements in large language model (LLM)-powered agents have shown that collective intelligence can significantly outperform individual capabilities, largely attributed to the meticulously designed inter-agent communication topologies. Though impressive in performance, existing multi-agent pipelines inherently introduce substantial token overhead, as well as increased economic costs, which pose challenges for their large-scale deployments. In response to this challenge, we propose an economical, simple, and robust multi-agent communication framework, termed $\texttt{AgentPrune}$, which can seamlessly integrate into mainstream multi-agent systems and prunes redundant or even malicious communication messages. Technically, $\texttt{AgentPrune}$ is the first to identify and formally define the \textit{communication redundancy} issue present in current LLM-based multi-agent pipelines, and efficiently performs one-shot pruning on the spatial-temporal message-passing graph, yielding a token-economic and high-performing communication topology. Extensive experiments across six benchmarks demonstrate that $\texttt{AgentPrune}$ \textbf{(I)} achieves comparable results as state-of-the-art topologies at merely $\$5.6$ cost compared to their $\$43.7$, \textbf{(II)} integrates seamlessly into existing multi-agent frameworks with $28.1\%\sim72.8\%\downarrow$ token reduction, and \textbf{(III)} successfully defend against two types of agent-based adversarial attacks with $3.5\%\sim10.8\%\uparrow$ performance boost.

☆ Dual Active Learning for Reinforcement Learning from Human Feedback

Pangpang Liu, Chengchun Shi, Will Wei Sun

Aligning large language models (LLMs) with human preferences is critical to recent advances in generative artificial intelligence. Reinforcement learning from human feedback (RLHF) is widely applied to achieve this objective. A key step in RLHF is to learn the reward function from human feedback. However, human feedback is costly and time-consuming, making it essential to collect high-quality conversation data for human teachers to label. Additionally, different human teachers have different levels of expertise. It is thus critical to query the most appropriate teacher for their opinions. In this paper, we use offline reinforcement learning (RL) to formulate the alignment problem. Motivated by the idea of $D$-optimal design, we first propose a dual active reward learning algorithm for the simultaneous selection of conversations and teachers. Next, we apply pessimistic RL to solve the alignment problem, based on the learned reward estimator. Theoretically, we show that the reward estimator obtained through our proposed adaptive selection strategy achieves minimal generalized variance asymptotically, and prove that the sub-optimality of our pessimistic policy scales as $O(1/\sqrt{T})$ with a given sample budget $T$. Through simulations and experiments on LLMs, we demonstrate the effectiveness of our algorithm and its superiority over state-of-the-arts.

☆ Dynamic Gradient Alignment for Online Data Mixing

Simin Fan, David Grangier, Pierre Ablin

☆ Efficient learning of differential network in multi-source non-paranormal graphical models

Mojtaba Nikahd, Seyed Abolfazl Motahari

This paper addresses learning of sparse structural changes or differential network between two classes of non-paranormal graphical models. We assume a multi-source and heterogeneous dataset is available for each class, where the covariance matrices are identical for all non-paranormal graphical models. The differential network, which are encoded by the difference precision matrix, can then be decoded by optimizing a lasso penalized D-trace loss function. To this aim, an efficient approach is proposed that outputs the exact solution path, outperforming the previous methods that only sample from the solution path in pre-selected regularization parameters. Notably, our proposed method has low computational complexity, especially when the differential network are sparse. Our simulations on synthetic data demonstrate a superior performance for our strategy in terms of speed and accuracy compared to an existing method. Moreover, our strategy in combining datasets from multiple sources is shown to be very effective in inferring differential network in real-world problems. This is backed by our experimental results on drug resistance in tumor cancers. In the latter case, our strategy outputs important genes for drug resistance which are already confirmed by various independent studies.

☆ Stochastic variance-reduced Gaussian variational inference on the Bures-Wasserstein manifold

Hoang Phuc Hau Luu, Hanlin Yu, Bernardo Williams, Marcelo Hartmann, Arto Klami

Optimization in the Bures-Wasserstein space has been gaining popularity in the machine learning community since it draws connections between variational inference and Wasserstein gradient flows. The variational inference objective function of Kullback-Leibler divergence can be written as the sum of the negative entropy and the potential energy, making forward-backward Euler the method of choice. Notably, the backward step admits a closed-form solution in this case, facilitating the practicality of the scheme. However, the forward step is no longer exact since the Bures-Wasserstein gradient of the potential energy involves "intractable" expectations. Recent approaches propose using the Monte Carlo method -- in practice a single-sample estimator -- to approximate these terms, resulting in high variance and poor performance. We propose a novel variance-reduced estimator based on the principle of control variates. We theoretically show that this estimator has a smaller variance than the Monte-Carlo estimator in scenarios of interest. We also prove that variance reduction helps improve the optimization bounds of the current analysis. We demonstrate that the proposed estimator gains order-of-magnitude improvements over the previous Bures-Wasserstein methods.

☆ Encryption-Friendly LLM Architecture

Donghwan Rho, Taeseong Kim, Minje Park, Jung Woo Kim, Hyunsik Chae, Jung Hee Cheon, Ernest K. Ryu

Large language models (LLMs) offer personalized responses based on user interactions, but this use case raises serious privacy concerns. Homomorphic encryption (HE) is a cryptographic protocol supporting arithmetic computations in encrypted states and provides a potential solution for privacy-preserving machine learning (PPML). However, the computational intensity of transformers poses challenges for applying HE to LLMs. In this work, we propose a modified HE-friendly transformer architecture with an emphasis on inference following personalized (private) fine-tuning. Utilizing LoRA fine-tuning and Gaussian kernels, we achieve significant computational speedups -- 6.94x for fine-tuning and 2.3x for inference -- while maintaining performance comparable to plaintext models. Our findings provide a viable proof of concept for offering privacy-preserving LLM services in areas where data protection is crucial.

comment: 27 pages

☆ Cross-Embodiment Dexterous Grasping with Reinforcement Learning

Haoqi Yuan, Bohan Zhou, Yuhui Fu, Zongqing Lu

Dexterous hands exhibit significant potential for complex real-world grasping tasks. While recent studies have primarily focused on learning policies for specific robotic hands, the development of a universal policy that controls diverse dexterous hands remains largely unexplored. In this work, we study the learning of cross-embodiment dexterous grasping policies using reinforcement learning (RL). Inspired by the capability of human hands to control various dexterous hands through teleoperation, we propose a universal action space based on the human hand's eigengrasps. The policy outputs eigengrasp actions that are then converted into specific joint actions for each robot hand through a retargeting mapping. We simplify the robot hand's proprioception to include only the positions of fingertips and the palm, offering a unified observation space across different robot hands. Our approach demonstrates an 80% success rate in grasping objects from the YCB dataset across four distinct embodiments using a single vision-based policy. Additionally, our policy exhibits zero-shot generalization to two previously unseen embodiments and significant improvement in efficient finetuning. For further details and videos, visit our project page https://sites.google.com/view/crossdex.

☆ Temporal Predictive Coding for Gradient Compression in Distributed Learning

Adrian Edin, Zheng Chen, Michel Kieffer, Mikael Johansson

comment: 8 pages, 3 figures, presented at the 60th Allerton conference on Communication, Control, and Computing

☆ Learning Diverse Bimanual Dexterous Manipulation Skills from Human Demonstrations

Bohan Zhou, Haoqi Yuan, Yuhui Fu, Zongqing Lu

Bimanual dexterous manipulation is a critical yet underexplored area in robotics. Its high-dimensional action space and inherent task complexity present significant challenges for policy learning, and the limited task diversity in existing benchmarks hinders general-purpose skill development. Existing approaches largely depend on reinforcement learning, often constrained by intricately designed reward functions tailored to a narrow set of tasks. In this work, we present a novel approach for efficiently learning diverse bimanual dexterous skills from abundant human demonstrations. Specifically, we introduce BiDexHD, a framework that unifies task construction from existing bimanual datasets and employs teacher-student policy learning to address all tasks. The teacher learns state-based policies using a general two-stage reward function across tasks with shared behaviors, while the student distills the learned multi-task policies into a vision-based policy. With BiDexHD, scalable learning of numerous bimanual dexterous skills from auto-constructed tasks becomes feasible, offering promising advances toward universal bimanual dexterous manipulation. Our empirical evaluation on the TACO dataset, spanning 141 tasks across six categories, demonstrates a task fulfillment rate of 74.59% on trained tasks and 51.07% on unseen tasks, showcasing the effectiveness and competitive zero-shot generalization capabilities of BiDexHD. For videos and more information, visit our project page https://sites.google.com/view/bidexhd.

☆ Online Convex Optimization with a Separation Oracle

Zakaria Mhammedi

In this paper, we introduce a new projection-free algorithm for Online Convex Optimization (OCO) with a state-of-the-art regret guarantee among separation-based algorithms. Existing projection-free methods based on the classical Frank-Wolfe algorithm achieve a suboptimal regret bound of $O(T^{3/4})$, while more recent separation-based approaches guarantee a regret bound of $O(\kappa \sqrt{T})$, where $\kappa$ denotes the asphericity of the feasible set, defined as the ratio of the radii of the containing and contained balls. However, for ill-conditioned sets, $\kappa$ can be arbitrarily large, potentially leading to poor performance. Our algorithm achieves a regret bound of $\tilde{O}(\sqrt{dT} + \kappa d)$, while requiring only $\tilde{O}(1)$ calls to a separation oracle per round. Crucially, the main term in the bound, $\tilde{O}(\sqrt{d T})$, is independent of $\kappa$, addressing the limitations of previous methods. Additionally, as a by-product of our analysis, we recover the $O(\kappa \sqrt{T})$ regret bound of existing OCO algorithms with a more straightforward analysis and improve the regret bound for projection-free online exp-concave optimization. Finally, for constrained stochastic convex optimization, we achieve a state-of-the-art convergence rate of $\tilde{O}(\sigma/\sqrt{T} + \kappa d/T)$, where $\sigma$ represents the noise in the stochastic gradients, while requiring only $\tilde{O}(1)$ calls to a separation oracle per iteration.

☆ Efficient Residual Learning with Mixture-of-Experts for Universal Dexterous Grasping

Ziye Huang, Haoqi Yuan, Yuhui Fu, Zongqing Lu

Universal dexterous grasping across diverse objects presents a fundamental yet formidable challenge in robot learning. Existing approaches using reinforcement learning (RL) to develop policies on extensive object datasets face critical limitations, including complex curriculum design for multi-task learning and limited generalization to unseen objects. To overcome these challenges, we introduce ResDex, a novel approach that integrates residual policy learning with a mixture-of-experts (MoE) framework. ResDex is distinguished by its use of geometry-unaware base policies that are efficiently acquired on individual objects and capable of generalizing across a wide range of unseen objects. Our MoE framework incorporates several base policies to facilitate diverse grasping styles suitable for various objects. By learning residual actions alongside weights that combine these base policies, ResDex enables efficient multi-task RL for universal dexterous grasping. ResDex achieves state-of-the-art performance on the DexGraspNet dataset comprising 3,200 objects with an 88.8% success rate. It exhibits no generalization gap with unseen objects and demonstrates superior training efficiency, mastering all tasks within only 12 hours on a single GPU.

☆ Meta-Models: An Architecture for Decoding LLM Behaviors Through Interpreted Embeddings and Natural Language

Anthony Costarelli, Mat Allen, Severin Field, Joshua Clymer

As Large Language Models (LLMs) become increasingly integrated into our daily lives, the potential harms from deceptive behavior underlie the need for faithfully interpreting their decision-making. While traditional probing methods have shown some effectiveness, they remain best for narrowly scoped tasks while more comprehensive explanations are still necessary. To this end, we investigate meta-models-an architecture using a "meta-model" that takes activations from an "input-model" and answers natural language questions about the input-model's behaviors. We evaluate the meta-model's ability to generalize by training them on selected task types and assessing their out-of-distribution performance in deceptive scenarios. Our findings show that meta-models generalize well to out-of-distribution tasks and point towards opportunities for future research in this area.

comment: 11 pages, 2 figures

☆ Towards a Theoretical Understanding of Memorization in Diffusion Models

Yunhao Chen, Xingjun Ma, Difan Zou, Yu-Gang Jiang

comment: arXiv admin note: text overlap with arXiv:2406.12752

☆ Quantifying User Coherence: A Unified Framework for Cross-Domain Recommendation Analysis

Michaël Soumm, Alexandre Fournier-Montgieux, Adrian Popescu, Bertrand Delezoide

The effectiveness of Recommender Systems (RS) is closely tied to the quality and distinctiveness of user profiles, yet despite many advancements in raw performance, the sensitivity of RS to user profile quality remains under-researched. This paper introduces novel information-theoretic measures for understanding recommender systems: a "surprise" measure quantifying users' deviations from popular choices, and a "conditional surprise" measure capturing user interaction coherence. We evaluate 7 recommendation algorithms across 9 datasets, revealing the relationships between our measures and standard performance metrics. Using a rigorous statistical framework, our analysis quantifies how much user profile density and information measures impact algorithm performance across domains. By segmenting users based on these measures, we achieve improved performance with reduced data and show that simpler algorithms can match complex ones for low-coherence users. Additionally, we employ our measures to analyze how well different recommendation algorithms maintain the coherence and diversity of user preferences in their predictions, providing insights into algorithm behavior. This work advances the theoretical understanding of user behavior and practical heuristics for personalized recommendation systems, promoting more efficient and adaptive architectures.

☆ Personalized Federated Learning for Generative AI-Assisted Semantic Communications

Yubo Peng, Feibo Jiang, Li Dong, Kezhi Wang, Kun Yang

☆ Clinnova Federated Learning Proof of Concept: Key Takeaways from a Cross-border Collaboration

Julia Alekseenko, Bram Stieltjes, Michael Bach, Melanie Boerries, Oliver Opitz, Alexandros Karargyris, Nicolas Padoy

☆ Learning K-U-Net with constant complexity: An Application to time series forecasting

Jiang You, Arben Cela, René Natowicz, Jacob Ouanounou, Patrick Siarry

Training deep models for time series forecasting is a critical task with an inherent challenge of time complexity. While current methods generally ensure linear time complexity, our observations on temporal redundancy show that high-level features are learned 98.44\% slower than low-level features. To address this issue, we introduce a new exponentially weighted stochastic gradient descent algorithm designed to achieve constant time complexity in deep learning models. We prove that the theoretical complexity of this learning method is constant. Evaluation of this method on Kernel U-Net (K-U-Net) on synthetic datasets shows a significant reduction in complexity while improving the accuracy of the test set.

☆ Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization

Mingyang Wang, Lukas Lange, Heike Adel, Jannik Strötgen, Hinrich Schütze

☆ Predictive Attractor Models NeurIPS 2024

Ramy Mounir, Sudeep Sarkar

comment: Accepted to NeurIPS 2024

☆ LLM-Pilot: Characterize and Optimize Performance of your LLM Inference Services SC '24

Małgorzata Łazuka, Andreea Anghel, Thomas Parnell

comment: Accepted to the International Conference for High Performance Computing, Networking, Storage and Analysis (SC '24)

☆ PnP-Flow: Plug-and-Play Image Restoration with Flow Matching

Ségolène Martin, Anne Gagneux, Paul Hagemann, Gabriele Steidl

☆ MenakBERT -- Hebrew Diacriticizer SC

Ido Cohen, Jacob Gidron, Idan Pinto

comment: Published at ISCOL2022 as a poster

☆ Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models

Seyedmorteza Sadat, Otmar Hilliges, Romann M. Weber

☆ An Online Feasible Point Method for Benign Generalized Nash Equilibrium Problems

Sarah Sachs, Hedi Hadiji, Tim van Erven, Mathias Staudigl

We consider a repeatedly played generalized Nash equilibrium game. This induces a multi-agent online learning problem with joint constraints. An important challenge in this setting is that the feasible set for each agent depends on the simultaneous moves of the other agents and, therefore, varies over time. As a consequence, the agents face time-varying constraints, which are not adversarial but rather endogenous to the system. Prior work in this setting focused on convergence to a feasible solution in the limit via integrating the constraints in the objective as a penalty function. However, no existing work can guarantee that the constraints are satisfied for all iterations while simultaneously guaranteeing convergence to a generalized Nash equilibrium. This is a problem of fundamental theoretical interest and practical relevance. In this work, we introduce a new online feasible point method. Under the assumption that limited communication between the agents is allowed, this method guarantees feasibility. We identify the class of benign generalized Nash equilibrium problems, for which the convergence of our method to the equilibrium is guaranteed. We set this class of benign generalized Nash equilibrium games in context with existing definitions and illustrate our method with examples.

☆ Parameter Competition Balancing for Model Merging NeurIPS2024

Guodong Du, Junlin Lee, Jing Li, Runhua Jiang, Yifei Guo, Shuyang Yu, Hanting Liu, Sim Kuan Goh, Ho-Kin Tang, Daojing He, Min Zhang

comment: Accepted by NeurIPS2024

☆ Online Multi-Label Classification under Noisy and Changing Label Distribution

Yizhang Zou, Xuegang Hu, Peipei Li, Jun Hu, You Wu

Multi-label data stream usually contains noisy labels in the real-world applications, namely occuring in both relevant and irrelevant labels. However, existing online multi-label classification methods are mostly limited in terms of label quality and fail to deal with the case of noisy labels. On the other hand, the ground-truth label distribution may vary with the time changing, which is hidden in the observed noisy label distribution and difficult to track, posing a major challenge for concept drift adaptation. Motivated by this, we propose an online multi-label classification algorithm under Noisy and Changing Label Distribution (NCLD). The convex objective is designed to simultaneously model the label scoring and the label ranking for high accuracy, whose robustness to NCLD benefits from three novel works: 1) The local feature graph is used to reconstruct the label scores jointly with the observed labels, and an unbiased ranking loss is derived and applied to learn reliable ranking information. 2) By detecting the difference between two adjacent chunks with the unbiased label cardinality, we identify the change in the ground-truth label distribution and reset the ranking or all information learned from the past to match the new distribution. 3) Efficient and accurate updating is achieved based on the updating rule derived from the closed-form optimal model solution. Finally, empirical experimental results validate the effectiveness of our method in classifying instances under NCLD.

☆ MANTRA: The Manifold Triangulations Assemblage

Rubén Ballester, Ernst Röell, Daniel Bin Schmid, Mathieu Alain, Sergio Escalera, Carles Casacuberta, Bastian Rieck

The rising interest in leveraging higher-order interactions present in complex systems has led to a surge in more expressive models exploiting high-order structures in the data, especially in topological deep learning (TDL), which designs neural networks on high-order domains such as simplicial complexes. However, progress in this field is hindered by the scarcity of datasets for benchmarking these architectures. To address this gap, we introduce MANTRA, the first large-scale, diverse, and intrinsically high order dataset for benchmarking high-order models, comprising over 43,000 and 249,000 triangulations of surfaces and three-dimensional manifolds, respectively. With MANTRA, we assess several graph- and simplicial complex-based models on three topological classification tasks. We demonstrate that while simplicial complex-based neural networks generally outperform their graph-based counterparts in capturing simple topological invariants, they also struggle, suggesting a rethink of TDL. Thus, MANTRA serves as a benchmark for assessing and advancing topological methods, leading the way for more effective high-order models.

comment: 26 pages, 2 figures, 22 tables

☆ Diffusion Meets Options: Hierarchical Generative Skill Composition for Temporally-Extended Tasks

Zeyu Feng, Hao Luan, Kevin Yuchen Ma, Harold Soh

Safe and successful deployment of robots requires not only the ability to generate complex plans but also the capacity to frequently replan and correct execution errors. This paper addresses the challenge of long-horizon trajectory planning under temporally extended objectives in a receding horizon manner. To this end, we propose DOPPLER, a data-driven hierarchical framework that generates and updates plans based on instruction specified by linear temporal logic (LTL). Our method decomposes temporal tasks into chain of options with hierarchical reinforcement learning from offline non-expert datasets. It leverages diffusion models to generate options with low-level actions. We devise a determinantal-guided posterior sampling technique during batch generation, which improves the speed and diversity of diffusion generated options, leading to more efficient querying. Experiments on robot navigation and manipulation tasks demonstrate that DOPPLER can generate sequences of trajectories that progressively satisfy the specified formulae for obstacle avoidance and sequential visitation. Demonstration videos are available online at: https://philiptheother.github.io/doppler/.

☆ BiSSL: Bilevel Optimization for Self-Supervised Pre-Training and Fine-Tuning

Gustav Wagner Zakarias, Lars Kai Hansen, Zheng-Hua Tan

In this work, we present BiSSL, a first-of-its-kind training framework that introduces bilevel optimization to enhance the alignment between the pretext pre-training and downstream fine-tuning stages in self-supervised learning. BiSSL formulates the pretext and downstream task objectives as the lower- and upper-level objectives in a bilevel optimization problem and serves as an intermediate training stage within the self-supervised learning pipeline. By more explicitly modeling the interdependence of these training stages, BiSSL facilitates enhanced information sharing between them, ultimately leading to a backbone parameter initialization that is better suited for the downstream task. We propose a training algorithm that alternates between optimizing the two objectives defined in BiSSL. Using a ResNet-18 backbone pre-trained with SimCLR on the STL10 dataset, we demonstrate that our proposed framework consistently achieves improved or competitive classification accuracies across various downstream image classification datasets compared to the conventional self-supervised learning pipeline. Qualitative analyses of the backbone features further suggest that BiSSL enhances the alignment of downstream features in the backbone prior to fine-tuning.

♻ ☆ CMP: Cooperative Motion Prediction with Multi-Agent Communication

Zehao Wang, Yuping Wang, Zhuoyuan Wu, Hengbo Ma, Zhaowei Li, Hang Qiu, Jiachen Li

comment: Project website: https://cmp-cooperative-prediction.github.io/

♻ ☆ Accelerating Training with Neuron Interaction and Nowcasting Networks

Boris Knyazev, Abhinav Moudgil, Guillaume Lajoie, Eugene Belilovsky, Simon Lacoste-Julien

Neural network training can be accelerated when a learnable update rule is used in lieu of classic adaptive optimizers (e.g. Adam). However, learnable update rules can be costly and unstable to train and use. Recently, Jang et al. (2023) proposed a simpler approach to accelerate training based on weight nowcaster networks (WNNs). In their approach, Adam is used for most of the optimization steps and periodically, only every few steps, a WNN nowcasts (predicts near future) parameters. We improve WNNs by proposing neuron interaction and nowcasting (NiNo) networks. In contrast to WNNs, NiNo leverages neuron connectivity and graph neural networks to more accurately nowcast parameters. We further show that in some networks, such as Transformers, modeling neuron connectivity accurately is challenging. We address this and other limitations, which allows NiNo to accelerate Adam training by up to 50% in vision and language tasks.

comment: added Llama3-based results and other updates, code is https://github.com/SamsungSAILMontreal/nino

♻ ☆ LML-DAP: Language Model Learning a Dataset for Data-Augmented Prediction

Praneeth Vadlapati

comment: Updated title, abstract, and images

♻ ☆ On Training Data Influence of GPT Models EMNLP 2024

Yekun Chai, Qingyi Liu, Shuohuan Wang, Yu Sun, Qiwei Peng, Hua Wu

comment: EMNLP 2024

♻ ☆ Preble: Efficient Distributed Prompt Scheduling for LLM Serving

Vikranth Srivatsa, Zijian He, Reyna Abhyankar, Dongming Li, Yiying Zhang

♻ ☆ E(n) Equivariant Topological Neural Networks

Claudio Battiloro, Ege Karaismailoğlu, Mauricio Tec, George Dasoulas, Michelle Audirac, Francesca Dominici

Graph neural networks excel at modeling pairwise interactions, but they cannot flexibly accommodate higher-order interactions and features. Topological deep learning (TDL) has emerged recently as a promising tool for addressing this issue. TDL enables the principled modeling of arbitrary multi-way, hierarchical higher-order interactions by operating on combinatorial topological spaces, such as simplicial or cell complexes, instead of graphs. However, little is known about how to leverage geometric features such as positions and velocities for TDL. This paper introduces E(n)-Equivariant Topological Neural Networks (ETNNs), which are E(n)-equivariant message-passing networks operating on combinatorial complexes, formal objects unifying graphs, hypergraphs, simplicial, path, and cell complexes. ETNNs incorporate geometric node features while respecting rotation, reflection, and translation equivariance. Moreover, ETNNs are natively ready for settings with heterogeneous interactions. We provide a theoretical analysis to show the improved expressiveness of ETNNs over architectures for geometric graphs. We also show how E(n)-equivariant variants of TDL models can be directly derived from our framework. The broad applicability of ETNNs is demonstrated through two tasks of vastly different scales: i) molecular property prediction on the QM9 benchmark and ii) land-use regression for hyper-local estimation of air pollution with multi-resolution irregular geospatial data. The results indicate that ETNNs are an effective tool for learning from diverse types of richly structured data, as they match or surpass SotA equivariant TDL models with a significantly smaller computational burden, thus highlighting the benefits of a principled geometric inductive bias.

comment: 41 pages, 11 figures, 12 tables

♻ ☆ Unichain and Aperiodicity are Sufficient for Asymptotic Optimality of Average-Reward Restless Bandits

Yige Hong, Qiaomin Xie, Yudong Chen, Weina Wang

We consider the infinite-horizon, average-reward restless bandit problem in discrete time. We propose a new class of policies that are designed to drive a progressively larger subset of arms toward the optimal distribution. We show that our policies are asymptotically optimal with an $O(1/\sqrt{N})$ optimality gap for an $N$-armed problem, assuming only a unichain and aperiodicity assumption. Our approach departs from most existing work that focuses on index or priority policies, which rely on the Global Attractor Property (GAP) to guarantee convergence to the optimum, or a recently developed simulation-based policy, which requires a Synchronization Assumption (SA).

comment: 58 pages, 14 figures. This version includes a restructured main result section and new experiments

♻ ☆ Lookback Lens: Detecting and Mitigating Contextual Hallucinations in Large Language Models Using Only Attention Maps EMNLP 2024

Yung-Sung Chuang, Linlu Qiu, Cheng-Yu Hsieh, Ranjay Krishna, Yoon Kim, James Glass

comment: EMNLP 2024 main conference long paper. The source code is available at https://github.com/voidism/Lookback-Lens

♻ ☆ VideoPhy: Evaluating Physical Commonsense for Video Generation

Hritik Bansal, Zongyu Lin, Tianyi Xie, Zeshun Zong, Michal Yarom, Yonatan Bitton, Chenfanfu Jiang, Yizhou Sun, Kai-Wei Chang, Aditya Grover

comment: 43 pages, 29 figures, 12 tables. Added CogVideo and Dream Machine in v2

♻ ☆ Collaborative learning of common latent representations in routinely collected multivariate ICU physiological signals ICASSP

Hollan Haule, Ian Piper, Patricia Jones, Tsz-Yan Milly Lo, Javier Escudero

In Intensive Care Units (ICU), the abundance of multivariate time series presents an opportunity for machine learning (ML) to enhance patient phenotyping. In contrast to previous research focused on electronic health records (EHR), here we propose an ML approach for phenotyping using routinely collected physiological time series data. Our new algorithm integrates Long Short-Term Memory (LSTM) networks with collaborative filtering concepts to identify common physiological states across patients. Tested on real-world ICU clinical data for intracranial hypertension (IH) detection in patients with brain injury, our method achieved an area under the curve (AUC) of 0.889 and average precision (AP) of 0.725. Moreover, our algorithm outperforms autoencoders in learning more structured latent representations of the physiological signals. These findings highlight the promise of our methodology for patient phenotyping, leveraging routinely collected multivariate time series to improve clinical care practices.

comment: Published in 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)

♻ ☆ Generalizing Medical Image Representations via Quaternion Wavelet Networks

Luigi Sigillo, Eleonora Grassucci, Aurelio Uncini, Danilo Comminiello

comment: This paper is currently under review

♻ ☆ On the Limited Generalization Capability of the Implicit Reward Model Induced by Direct Preference Optimization EMNLP

Yong Lin, Skyler Seto, Maartje ter Hoeve, Katherine Metcalf, Barry-John Theobald, Xuan Wang, Yizhe Zhang, Chen Huang, Tong Zhang

comment: 12 pages, 8 tables, 3 figures; Paper Accepted at EMNLP Findings 2024

♻ ☆ Jailbreaking LLMs with Arabic Transliteration and Arabizi EMNLP 2024

Mansour Al Ghanim, Saleh Almohaimeed, Mengxin Zheng, Yan Solihin, Qian Lou

comment: Accepted by EMNLP 2024

♻ ☆ Fair Allocation in Dynamic Mechanism Design NeurIPS

Alireza Fallah, Michael I. Jordan, Annie Ulichney

We consider a dynamic mechanism design problem where an auctioneer sells an indivisible good to groups of buyers in every round, for a total of $T$ rounds. The auctioneer aims to maximize their discounted overall revenue while adhering to a fairness constraint that guarantees a minimum average allocation for each group. We begin by studying the static case ($T=1$) and establish that the optimal mechanism involves two types of subsidization: one that increases the overall probability of allocation to all buyers, and another that favors the groups which otherwise have a lower probability of winning the item. We then extend our results to the dynamic case by characterizing a set of recursive functions that determine the optimal allocation and payments in each round. Notably, our results establish that in the dynamic case, the seller, on the one hand, commits to a participation bonus to incentivize truth-telling, and on the other hand, charges an entry fee for every round. Moreover, the optimal allocation once more involves subsidization, which its extent depends on the difference in future utilities for both the seller and buyers when allocating the item to one group versus the others. Finally, we present an approximation scheme to solve the recursive equations and determine an approximately optimal and fair allocation efficiently.

comment: A shorter conference version has been accepted at the Advances in Neural Information Processing Systems (NeurIPS) 2024

♻ ☆ Signature Isolation Forest

Marta Campi, Guillaume Staerman, Gareth W. Peters, Tomoko Matsui

Functional Isolation Forest (FIF) is a recent state-of-the-art Anomaly Detection (AD) algorithm designed for functional data. It relies on a tree partition procedure where an abnormality score is computed by projecting each curve observation on a drawn dictionary through a linear inner product. Such linear inner product and the dictionary are a priori choices that highly influence the algorithm's performances and might lead to unreliable results, particularly with complex datasets. This work addresses these challenges by introducing \textit{Signature Isolation Forest}, a novel AD algorithm class leveraging the rough path theory's signature transform. Our objective is to remove the constraints imposed by FIF through the proposition of two algorithms which specifically target the linearity of the FIF inner product and the choice of the dictionary. We provide several numerical experiments, including a real-world applications benchmark showing the relevance of our methods.

♻ ☆ DyGPrompt: Learning Feature and Time Prompts on Dynamic Graphs

Xingtong Yu, Zhenghao Liu, Yuan Fang, Xinming Zhang

Dynamic graphs capture evolving interactions between entities, such as in social networks, online learning platforms, and crowdsourcing projects. For dynamic graph modeling, dynamic graph neural networks (DGNNs) have emerged as a mainstream technique. However, they are generally pre-trained on the link prediction task, leaving a significant gap from the objectives of downstream tasks such as node classification. To bridge the gap, prompt-based learning has gained traction on graphs, but most existing efforts focus on static graphs, neglecting the evolution of dynamic graphs. In this paper, we propose DYGPROMPT, a novel pre-training and prompt learning framework for dynamic graph modeling. First, we design dual prompts to address the gap in both task objectives and temporal variations across pre-training and downstream tasks. Second, we recognize that node and time features mutually characterize each other, and propose dual condition-nets to model the evolving node-time patterns in downstream tasks. Finally, we thoroughly evaluate and analyze DYGPROMPT through extensive experiments on four public datasets.

comment: Under review

♻ ☆ Does Refusal Training in LLMs Generalize to the Past Tense?

Maksym Andriushchenko, Nicolas Flammarion

comment: Update in v3: o1-mini and o1-preview results (on top of GPT-4o and Claude 3.5 Sonnet added in v2). We provide code and jailbreak artifacts at https://github.com/tml-epfl/llm-past-tense

♻ ☆ Scalable Label Distribution Learning for Multi-Label Classification

Xingyu Zhao, Yuexuan An, Lei Qi, Xin Geng

Multi-label classification (MLC) refers to the problem of tagging a given instance with a set of relevant labels. Most existing MLC methods are based on the assumption that the correlation of two labels in each label pair is symmetric, which is violated in many real-world scenarios. Moreover, most existing methods design learning processes associated with the number of labels, which makes their computational complexity a bottleneck when scaling up to large-scale output space. To tackle these issues, we propose a novel method named Scalable Label Distribution Learning (SLDL) for multi-label classification which can describe different labels as distributions in a latent space, where the label correlation is asymmetric and the dimension is independent of the number of labels. Specifically, SLDL first converts labels into continuous distributions within a low-dimensional latent space and leverages the asymmetric metric to establish the correlation between different labels. Then, it learns the mapping from the feature space to the latent space, resulting in the computational complexity is no longer related to the number of labels. Finally, SLDL leverages a nearest-neighbor-based strategy to decode the latent representations and obtain the final predictions. Extensive experiments illustrate that SLDL achieves very competitive classification performances with little computational consumption.

♻ ☆ Foundations of Large Language Model Compression -- Part 1: Weight Quantization

Sean I. Young

comment: Preprint. 17 pages, 4 figures, 5 appendices

♻ ☆ NECOMIMI: Neural-Cognitive Multimodal EEG-informed Image Generation with Diffusion Models

Chi-Sheng Chen

NECOMIMI (NEural-COgnitive MultImodal EEG-Informed Image Generation with Diffusion Models) introduces a novel framework for generating images directly from EEG signals using advanced diffusion models. Unlike previous works that focused solely on EEG-image classification through contrastive learning, NECOMIMI extends this task to image generation. The proposed NERV EEG encoder demonstrates state-of-the-art (SoTA) performance across multiple zero-shot classification tasks, including 2-way, 4-way, and 200-way, and achieves top results in our newly proposed Category-based Assessment Table (CAT) Score, which evaluates the quality of EEG-generated images based on semantic concepts. A key discovery of this work is that the model tends to generate abstract or generalized images, such as landscapes, rather than specific objects, highlighting the inherent challenges of translating noisy and low-resolution EEG data into detailed visual outputs. Additionally, we introduce the CAT Score as a new metric tailored for EEG-to-image evaluation and establish a benchmark on the ThingsEEG dataset. This study underscores the potential of EEG-to-image generation while revealing the complexities and challenges that remain in bridging neural activity with visual representation.

♻ ☆ EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage

Zeyi Liao, Lingbo Mo, Chejian Xu, Mintong Kang, Jiawei Zhang, Chaowei Xiao, Yuan Tian, Bo Li, Huan Sun

comment: 29 pages

♻ ☆ Graph Diffusion Transformers for Multi-Conditional Molecular Generation NeurIPS 2024

Gang Liu, Jiaxin Xu, Tengfei Luo, Meng Jiang

Inverse molecular design with diffusion models holds great potential for advancements in material and drug discovery. Despite success in unconditional molecular generation, integrating multiple properties such as synthetic score and gas permeability as condition constraints into diffusion models remains unexplored. We present the Graph Diffusion Transformer (Graph DiT) for multi-conditional molecular generation. Graph DiT integrates an encoder to learn numerical and categorical property representations with the Transformer-based denoiser. Unlike previous graph diffusion models that add noise separately on the atoms and bonds in the forward diffusion process, Graph DiT is trained with a novel graph-dependent noise model for accurate estimation of graph-related noise in molecules. We extensively validate Graph DiT for multi-conditional polymer and small molecule generation. Results demonstrate the superiority of Graph DiT across nine metrics from distribution learning to condition control for molecular properties. A polymer inverse design task for gas separation with feedback from domain experts further demonstrates its practical utility.

comment: Accepted by NeurIPS 2024 (Oral). 21 pages, 11 figures, 8 tables

♻ ☆ Sample and Oracle Efficient Reinforcement Learning for MDPs with Linearly-Realizable Value Functions

Zakaria Mhammedi

Designing sample-efficient and computationally feasible reinforcement learning (RL) algorithms is particularly challenging in environments with large or infinite state and action spaces. In this paper, we advance this effort by presenting an efficient algorithm for Markov Decision Processes (MDPs) where the state-action value function of any policy is linear in a given feature map. This challenging setting can model environments with infinite states and actions, strictly generalizes classic linear MDPs, and currently lacks a computationally efficient algorithm under online access to the MDP. Specifically, we introduce a new RL algorithm that efficiently finds a near-optimal policy in this setting, using a number of episodes and calls to a cost-sensitive classification (CSC) oracle that are both polynomial in the problem parameters. Notably, our CSC oracle can be efficiently implemented when the feature dimension is constant, representing a clear improvement over state-of-the-art methods, which require solving non-convex problems with horizon-many variables and can incur computational costs that are exponential in the horizon.

♻ ☆ Adjusted Expected Improvement for Cumulative Regret Minimization in Noisy Bayesian Optimization

Shouri Hu, Haowei Wang, Zhongxiang Dai, Bryan Kian Hsiang Low, Szu Hui Ng

The expected improvement (EI) is one of the most popular acquisition functions for Bayesian optimization (BO) and has demonstrated good empirical performances in many applications for the minimization of simple regret. However, under the evaluation metric of cumulative regret, the performance of EI may not be competitive, and its existing theoretical regret upper bound still has room for improvement. To adapt the EI for better performance under cumulative regret, we introduce a novel quantity called the evaluation cost which is compared against the acquisition function, and with this, develop the expected improvement-cost (EIC) algorithm. In each iteration of EIC, a new point with the largest acquisition function value is sampled, only if that value exceeds its evaluation cost. If none meets this criteria, the current best point is resampled. This evaluation cost quantifies the potential downside of sampling a point, which is important under the cumulative regret metric as the objective function value in every iteration affects the performance measure. We establish in theory a high-probability regret upper bound of EIC based on the maximum information gain, which is tighter than the bound of existing EI-based algorithms. It is also comparable to the regret bound of other popular BO algorithms such as Thompson sampling (GP-TS) and upper confidence bound (GP-UCB). We further perform experiments to illustrate the improvement of EIC over several popular BO algorithms.

♻ ☆ A Methodological Report on Anomaly Detection on Dynamic Knowledge Graphs

Xiaohua Lu, Leshanshui Yang

In this paper, we explore different approaches to anomaly detection on dynamic knowledge graphs, specifically in a microservices environment for Kubernetes applications. Our approach explores three dynamic knowledge graph representations: sequential data, one-hop graph structure, and two-hop graph structure, with each representation incorporating increasingly complex structural information. Each phase includes different machine learning and deep learning models. We empirically analyse their performance and propose an approach based on ensemble learning of these models. Our approach significantly outperforms the baseline on the ISWC 2024 Dynamic Knowledge Graph Anomaly Detection dataset, providing a robust solution for anomaly detection in dynamic complex data.

♻ ☆ A Deep Generative Learning Approach for Two-stage Adaptive Robust Optimization

Aron Brenner, Rahman Khorramfar, Jennifer Sun, Saurabh Amin

Two-stage adaptive robust optimization (ARO) is a powerful approach for planning under uncertainty, balancing first-stage decisions with recourse decisions made after uncertainty is realized. To account for uncertainty, modelers typically define a simple uncertainty set over which potential outcomes are considered. However, classical methods for defining these sets unintentionally capture a wide range of unrealistic outcomes, resulting in overly-conservative and costly planning in anticipation of unlikely contingencies. In this work, we introduce AGRO, a solution algorithm that performs adversarial generation for two-stage adaptive robust optimization using a variational autoencoder. AGRO generates high-dimensional contingencies that are simultaneously adversarial and realistic, improving the robustness of first-stage decisions at a lower planning cost than standard methods. To ensure generated contingencies lie in high-density regions of the uncertainty distribution, AGRO defines a tight uncertainty set as the image of "latent" uncertainty sets under the VAE decoding transformation. Projected gradient ascent is then used to maximize recourse costs over the latent uncertainty sets by leveraging differentiable optimization methods. We demonstrate the cost-efficiency of AGRO by applying it to both a synthetic production-distribution problem and a real-world power system expansion setting. We show that AGRO outperforms the standard column-and-constraint algorithm by up to 1.8% in production-distribution planning and up to 11.6% in power system expansion.

♻ ☆ PARAMANU-AYN: Pretrain from scratch or Continual Pretraining of LLMs for Legal Domain Adaptation?

Mitodru Niyogi, Arnab Bhattacharya

♻ ☆ Advantage Alignment Algorithms

Juan Agustin Duque, Milad Aghajohari, Tim Cooijmans, Razvan Ciuca, Tianyu Zhang, Gauthier Gidel, Aaron Courville

Artificially intelligent agents are increasingly being integrated into human decision-making: from large language model (LLM) assistants to autonomous vehicles. These systems often optimize their individual objective, leading to conflicts, particularly in general-sum games where naive reinforcement learning agents empirically converge to Pareto-suboptimal Nash equilibria. To address this issue, opponent shaping has emerged as a paradigm for finding socially beneficial equilibria in general-sum games. In this work, we introduce Advantage Alignment, a family of algorithms derived from first principles that perform opponent shaping efficiently and intuitively. We achieve this by aligning the advantages of interacting agents, increasing the probability of mutually beneficial actions when their interaction has been positive. We prove that existing opponent shaping methods implicitly perform Advantage Alignment. Compared to these methods, Advantage Alignment simplifies the mathematical formulation of opponent shaping, reduces the computational burden and extends to continuous action domains. We demonstrate the effectiveness of our algorithms across a range of social dilemmas, achieving state-of-the-art cooperation and robustness against exploitation.

comment: 25 Pages, 8 figures

♻ ☆ Fast Matrix Multiplications for Lookup Table-Quantized LLMs EMNLP 2024

Han Guo, William Brandon, Radostin Cholakov, Jonathan Ragan-Kelley, Eric P. Xing, Yoon Kim

comment: EMNLP 2024 (Findings)

♻ ☆ Nebula: A discourse aware Minecraft Builder EMNLP 2024

Akshay Chaturvedi, Kate Thompson, Nicholas Asher

comment: EMNLP 2024 Findings

♻ ☆ LongForm: Effective Instruction Tuning with Reverse Instructions EMNLP 2024

Abdullatif Köksal, Timo Schick, Anna Korhonen, Hinrich Schütze

comment: EMNLP 2024 Findings. This version extends the training with recent LLMs, evaluation with new metrics, and NLU tasks

♻ ☆ Residual-based Attention Physics-informed Neural Networks for Spatio-Temporal Ageing Assessment of Transformers Operated in Renewable Power Plants

Ibai Ramirez, Joel Pino, David Pardo, Mikel Sanz, Luis del Rio, Alvaro Ortiz, Kateryna Morozovska, Jose I. Aizpurua

Transformers are crucial for reliable and efficient power system operations, particularly in supporting the integration of renewable energy. Effective monitoring of transformer health is critical to maintain grid stability and performance. Thermal insulation ageing is a key transformer failure mode, which is generally tracked by monitoring the hotspot temperature (HST). However, HST measurement is complex, costly, and often estimated from indirect measurements. Existing HST models focus on space-agnostic thermal models, providing worst-case HST estimates. This article introduces a spatio-temporal model for transformer winding temperature and ageing estimation, which leverages physics-based partial differential equations (PDEs) with data-driven Neural Networks (NN) in a Physics Informed Neural Networks (PINNs) configuration to improve prediction accuracy and acquire spatio-temporal resolution. The computational accuracy of the PINN model is improved through the implementation of the Residual-Based Attention (PINN-RBA) scheme that accelerates the PINN model convergence. The PINN-RBA model is benchmarked against self-adaptive attention schemes and classical vanilla PINN configurations. For the first time, PINN based oil temperature predictions are used to estimate spatio-temporal transformer winding temperature values, validated through PDE numerical solution and fiber optic sensor measurements. Furthermore, the spatio-temporal transformer ageing model is inferred, which supports transformer health management decision-making. Results are validated with a distribution transformer operating on a floating photovoltaic power plant.

comment: 23 pages, 18 figures

♻ ☆ Embodied-RAG: General Non-parametric Embodied Memory for Retrieval and Generation

Quanting Xie, So Yeon Min, Tianyi Zhang, Aarav Bajaj, Ruslan Salakhutdinov, Matthew Johnson-Roberson, Yonatan Bisk

There is no limit to how much a robot might explore and learn, but all of that knowledge needs to be searchable and actionable. Within language research, retrieval augmented generation (RAG) has become the workhouse of large-scale non-parametric knowledge, however existing techniques do not directly transfer to the embodied domain, which is multimodal, data is highly correlated, and perception requires abstraction. To address these challenges, we introduce Embodied-RAG, a framework that enhances the foundational model of an embodied agent with a non-parametric memory system capable of autonomously constructing hierarchical knowledge for both navigation and language generation. Embodied-RAG handles a full range of spatial and semantic resolutions across diverse environments and query types, whether for a specific object or a holistic description of ambiance. At its core, Embodied-RAG's memory is structured as a semantic forest, storing language descriptions at varying levels of detail. This hierarchical organization allows the system to efficiently generate context-sensitive outputs across different robotic platforms. We demonstrate that Embodied-RAG effectively bridges RAG to the robotics domain, successfully handling over 200 explanation and navigation queries across 19 environments, highlighting its promise for general-purpose non-parametric system for embodied agents.

comment: Web: https://quanting-xie.github.io/Embodied-RAG-web/

♻ ☆ LDMol: Text-to-Molecule Diffusion Model with Structurally Informative Latent Space

Jinho Chang, Jong Chul Ye

With the emergence of diffusion models as the frontline of generative models, many researchers have proposed molecule generation techniques with conditional diffusion models. However, the unavoidable discreteness of a molecule makes it difficult for a diffusion model to connect raw data with highly complex conditions like natural language. To address this, we present a novel latent diffusion model dubbed LDMol for text-conditioned molecule generation. LDMol comprises a molecule autoencoder that produces a learnable and structurally informative feature space, and a natural language-conditioned latent diffusion model. In particular, recognizing that multiple SMILES notations can represent the same molecule, we employ a contrastive learning strategy to extract feature space that is aware of the unique characteristics of the molecule structure. LDMol outperforms the existing baselines on the text-to-molecule generation benchmark, suggesting a potential for diffusion models can outperform autoregressive models in text data generation with a better choice of the latent domain. Furthermore, we show that LDMol can be applied to downstream tasks such as molecule-to-text retrieval and text-guided molecule editing, demonstrating its versatility as a diffusion model.

♻ ☆ Learning an Actionable Discrete Diffusion Policy via Large-Scale Actionless Video Pre-Training NeurIPS 2024

Haoran He, Chenjia Bai, Ling Pan, Weinan Zhang, Bin Zhao, Xuelong Li

comment: Accepted by NeurIPS 2024. 24 pages

♻ ☆ On Sensitivity of Learning with Limited Labelled Data to the Effects of Randomness: Impact of Interactions and Systematic Choices EMNLP'24

Branislav Pecher, Ivan Srba, Maria Bielikova

comment: Accepted to the EMNLP'24 Main Conference

♻ ☆ AtomSurf : Surface Representation for Learning on Protein Structures

Vincent Mallet, Souhaib Attaiki, Yangyang Miao, Bruno Correia, Maks Ovsjanikov

While there has been significant progress in evaluating and comparing different representations for learning on protein data, the role of surface-based learning approaches remains not well-understood. In particular, there is a lack of direct and fair benchmark comparison between the best available surface-based learning methods against alternative representations such as graphs. Moreover, the few existing surface-based approaches either use surface information in isolation or, at best, perform global pooling between surface and graph-based architectures. In this work, we fill this gap by first adapting a state-of-the-art surface encoder for protein learning tasks. We then perform a direct and fair comparison of the resulting method against alternative approaches within the Atom3D benchmark, highlighting the limitations of pure surface-based learning. Finally, we propose an integrated approach, which allows learned feature sharing between graphs and surface representations on the level of nodes and vertices $\textit{across all layers}$. We demonstrate that the resulting architecture achieves state-of-the-art results on all tasks in the Atom3D benchmark, while adhering to the strict benchmark protocol, as well as more broadly on binding site identification and binding pocket classification. Furthermore, we use coarsened surfaces and optimize our approach for efficiency, making our tool competitive in training and inference time with existing techniques. Our code and data can be found online: $\texttt{github.com/Vincentx15/atomsurf}$

comment: 10 pages

♻ ☆ PowerPM: Foundation Model for Power Systems

Shihao Tu, Yupeng Zhang, Jing Zhang, Zhendong Fu, Yin Zhang, Yang Yang

The emergence of abundant electricity time series (ETS) data provides ample opportunities for various applications in the power systems, including demand-side management, grid stability, and consumer behavior analysis. Deep learning models have advanced ETS modeling by effectively capturing sequence dependence. Nevertheless, learning a generic representation of ETS data for various applications remains challenging due to the inherently complex hierarchical structure of ETS data. Moreover, ETS data exhibits intricate temporal dependencies and is suscepti ble to the influence of exogenous variables. Furthermore, different instances exhibit diverse electricity consumption behavior. In this paper, we propose a foundation model PowerPM to model ETS data, providing a large-scale, off-the-shelf model for power systems. PowerPM consists of a temporal encoder and a hierarchical encoder. The temporal encoder captures both temporal dependencies in ETS data, considering exogenous variables. The hierarchical encoder models the correlation between hierarchy. Furthermore, PowerPM leverages a novel self-supervised pretraining framework consisting of masked ETS modeling and dual-view contrastive learning, which enable PowerPM to capture temporal dependency within ETS windows and aware the discrepancy across ETS windows, providing two different perspectives to learn generic representation. Our experiments involve five real world scenario datasets, comprising private and public data. Through pre-training on massive ETS data, PowerPM achieves SOTA performance on diverse downstream tasks within the private dataset. Impressively, when transferred to the public datasets, PowerPM maintains its superiority, showcasing its remarkable generalization ability across various tasks and domains. Moreover, ablation studies, few-shot experiments provide additional evidence of the effectiveness of our model.

comment: 23 pages, 5 figures, 8 tables

♻ ☆ Solution of the Probabilistic Lambert Problem: Connections with Optimal Mass Transport, Schrödinger Bridge and Reaction-Diffusion PDEs

Alexis M. H. Teter, Iman Nodozi, Abhishek Halder

The Lambert problem originated in orbital mechanics. It concerns with determining the initial velocity for a boundary value problem involving the dynamical constraint due to gravitational potential with additional time horizon and endpoint position constraints. Its solution has application in transferring a spacecraft from a given initial to a given terminal position within prescribed flight time via velocity control. We consider a probabilistic variant of the Lambert problem where the knowledge of the endpoint constraints in position vectors are replaced by the knowledge of their respective joint probability density functions. We show that the Lambert problem with endpoint joint probability density constraints is a generalized optimal mass transport (OMT) problem, thereby connecting this classical astrodynamics problem with a burgeoning area of research in modern stochastic control and stochastic machine learning. This newfound connection allows us to rigorously establish the existence and uniqueness of solution for the probabilistic Lambert problem. The same connection also helps to numerically solve the probabilistic Lambert problem via diffusion regularization, i.e., by leveraging further connection of the OMT with the Schr\"odinger bridge problem (SBP). This also shows that the probabilistic Lambert problem with additive dynamic process noise is a generalized SBP, and can be solved numerically using the so-called Schr\"odinger factors, as we do in this work. Our analysis leads to solving a system of reaction-diffusion PDEs where the gravitational potential appears as the reaction rate.

♻ ☆ Source-Free Domain Adaptation Guided by Vision and Vision-Language Pre-Training ICCV

Wenyu Zhang, Li Shen, Chuan-Sheng Foo

comment: Extension of ICCV paper arXiv:2212.07585; Published at IJCV

♻ ☆ Symbolic State Partitioning for Reinforcement Learning

Mohsen Ghaffari, Mahsa Varshosaz, Einar Broch Johnsen, Andrzej Wąsowski

Tabular reinforcement learning methods cannot operate directly on continuous state spaces. One solution for this problem is to partition the state space. A good partitioning enables generalization during learning and more efficient exploitation of prior experiences. Consequently, the learning process becomes faster and produces more reliable policies. However, partitioning introduces approximation, which is particularly harmful in the presence of nonlinear relations between state components. An ideal partition should be as coarse as possible, while capturing the key structure of the state space for the given problem. This work extracts partitions from the environment dynamics by symbolic execution. We show that symbolic partitioning improves state space coverage with respect to environmental behavior and allows reinforcement learning to perform better for sparse rewards. We evaluate symbolic state space partitioning with respect to precision, scalability, learning agent performance and state space coverage for the learnt policies.

♻ ☆ miniCTX: Neural Theorem Proving with (Long-)Contexts

Jiewen Hu, Thomas Zhu, Sean Welleck

♻ ☆ A Causal Bayesian Network and Probabilistic Programming Based Reasoning Framework for Robot Manipulation Under Uncertainty ICRA 2025

Ricardo Cannizzaro, Michael Groom, Jonathan Routley, Robert Osazuwa Ness, Lars Kunze

Robot object manipulation in real-world environments is challenging because robot operation must be robust to a range of sensing, estimation, and actuation uncertainties to avoid potentially unsafe and costly mistakes that are a barrier to their adoption. In this paper, we propose a flexible and generalisable physics-informed causal Bayesian network (CBN) based framework for a robot to probabilistically reason about candidate manipulation actions, to enable robot decision-making robust to arbitrary robot system uncertainties -- the first of its kind to use a probabilistic programming language implementation. Using experiments in high-fidelity Gazebo simulation of an exemplar block stacking task, we demonstrate our framework's ability to: (1) predict manipulation outcomes with high accuracy (Pred Acc: 88.6%); and, (2) perform greedy next-best action selection with 94.2% task success rate. We also demonstrate our framework's suitability for real-world robot systems with a domestic robot. Thus, we show that by combining probabilistic causal modelling with physics simulations, we can make robot manipulation more robust to system uncertainties and hence more feasible for real-world applications. Further, our generalised reasoning framework can be used and extended for future robotics and causality research.

comment: 7 pages, 7 figures, submitted to the 2025 IEEE Conference on Robotics and Automation (ICRA 2025)

♻ ☆ Regret-Optimal Federated Transfer Learning for Kernel Regression with Applications in American Option Pricing

Xuwei Yang, Anastasis Kratsios, Florian Krach, Matheus Grasselli, Aurelien Lucchi

We propose an optimal iterative scheme for federated transfer learning, where a central planner has access to datasets ${\cal D}_1,\dots,{\cal D}_N$ for the same learning model $f_{\theta}$. Our objective is to minimize the cumulative deviation of the generated parameters $\{\theta_i(t)\}_{t=0}^T$ across all $T$ iterations from the specialized parameters $\theta^\star_{1},\ldots,\theta^\star_N$ obtained for each dataset, while respecting the loss function for the model $f_{\theta(T)}$ produced by the algorithm upon halting. We only allow for continual communication between each of the specialized models (nodes/agents) and the central planner (server), at each iteration (round). For the case where the model $f_{\theta}$ is a finite-rank kernel regression, we derive explicit updates for the regret-optimal algorithm. By leveraging symmetries within the regret-optimal algorithm, we further develop a nearly regret-optimal heuristic that runs with $\mathcal{O}(Np^2)$ fewer elementary operations, where $p$ is the dimension of the parameter space. Additionally, we investigate the adversarial robustness of the regret-optimal algorithm showing that an adversary which perturbs $q$ training pairs by at-most $\varepsilon>0$, across all training sets, cannot reduce the regret-optimal algorithm's regret by more than $\mathcal{O}(\varepsilon q \bar{N}^{1/2})$, where $\bar{N}$ is the aggregate number of training pairs. To validate our theoretical findings, we conduct numerical experiments in the context of American option pricing, utilizing a randomly generated finite-rank kernel.

comment: 51 pages, 2 figures

♻ ☆ Graph Chain-of-Thought: Augmenting Large Language Models by Reasoning on Graphs

Bowen Jin, Chulin Xie, Jiawei Zhang, Kashob Kumar Roy, Yu Zhang, Zheng Li, Ruirui Li, Xianfeng Tang, Suhang Wang, Yu Meng, Jiawei Han

comment: 21 pages. Code: https://github.com/PeterGriffinJin/Graph-CoT

♻ ☆ A Systematic Survey and Critical Review on Evaluating Large Language Models: Challenges, Limitations, and Recommendations EMNLP 2024

comment: Accepted at EMNLP 2024 (Main Conference)

♻ ☆ Forecasting Disease Progression with Parallel Hyperplanes in Longitudinal Retinal OCT MICCAI 2024

Arunava Chakravarty, Taha Emre, Dmitrii Lachinov, Antoine Rivail, Hendrik Scholl, Lars Fritsche, Sobha Sivaprasad, Daniel Rueckert, Andrew Lotery, Ursula Schmidt-Erfurth, Hrvoje Bogunović

comment: accepted in MICCAI 2024

♻ ☆ Large Language Models on Graphs: A Comprehensive Survey

Bowen Jin, Gang Liu, Chi Han, Meng Jiang, Heng Ji, Jiawei Han

comment: 25 pages

♻ ☆ Synthetic continued pretraining

Zitong Yang, Neil Band, Shuangping Li, Emmanuel Candès, Tatsunori Hashimoto

comment: Updated organization of experimental results and methods introduction. Released the dataset and model weights artifact

♻ ☆ Effective Heterogeneous Federated Learning via Efficient Hypernetwork-based Weight Generation

Yujin Shin, Kichang Lee, Sungmin Lee, You Rim Choi, Hyung-Sin Kim, JeongGil Ko

♻ ☆ Graph Sparsification via Mixture of Graphs

Guibin Zhang, Xiangguo Sun, Yanwei Yue, Chonghe Jiang, Kun Wang, Tianlong Chen, Shirui Pan

Graph Neural Networks (GNNs) have demonstrated superior performance across various graph learning tasks but face significant computational challenges when applied to large-scale graphs. One effective approach to mitigate these challenges is graph sparsification, which involves removing non-essential edges to reduce computational overhead. However, previous graph sparsification methods often rely on a single global sparsity setting and uniform pruning criteria, failing to provide customized sparsification schemes for each node's complex local context. In this paper, we introduce Mixture-of-Graphs (MoG), leveraging the concept of Mixture-of-Experts (MoE), to dynamically select tailored pruning solutions for each node. Specifically, MoG incorporates multiple sparsifier experts, each characterized by unique sparsity levels and pruning criteria, and selects the appropriate experts for each node. Subsequently, MoG performs a mixture of the sparse graphs produced by different experts on the Grassmann manifold to derive an optimal sparse graph. One notable property of MoG is its entirely local nature, as it depends on the specific circumstances of each individual node. Extensive experiments on four large-scale OGB datasets and two superpixel datasets, equipped with five GNN backbones, demonstrate that MoG (I) identifies subgraphs at higher sparsity levels ($8.67\%\sim 50.85\%$), with performance equal to or better than the dense graph, (II) achieves $1.47-2.62\times$ speedup in GNN inference with negligible performance drop, and (III) boosts ``top-student'' GNN performance ($1.02\%\uparrow$ on RevGNN+\textsc{ogbn-proteins} and $1.74\%\uparrow$ on DeeperGCN+\textsc{ogbg-ppa}).

♻ ☆ Functional Latent Dynamics for Irregularly Sampled Time Series Forecasting

Christian Klötergens, Vijaya Krishna Yalavarthi, Maximilian Stubbemann, Lars Schmidt-Thieme

Irregularly sampled time series with missing values are often observed in multiple real-world applications such as healthcare, climate and astronomy. They pose a significant challenge to standard deep learning models that operate only on fully observed and regularly sampled time series. In order to capture the continuous dynamics of the irregular time series, many models rely on solving an Ordinary Differential Equation (ODE) in the hidden state. These ODE-based models tend to perform slow and require large memory due to sequential operations and a complex ODE solver. As an alternative to complex ODE-based models, we propose a family of models called Functional Latent Dynamics (FLD). Instead of solving the ODE, we use simple curves which exist at all time points to specify the continuous latent state in the model. The coefficients of these curves are learned only from the observed values in the time series ignoring the missing values. Through extensive experiments, we demonstrate that FLD achieves better performance compared to the best ODE-based model while reducing the runtime and memory overhead. Specifically, FLD requires an order of magnitude less time to infer the forecasts compared to the best performing forecasting model.

♻ ☆ Hybrid Quantum-inspired Resnet and Densenet for Pattern Recognition

Andi Chen, Hua-Lei Yin, Zeng-Bing Chen, Shengjun Wu

In this paper, we propose two hybrid quantum-inspired neural networks with residual and dense connections respectively for pattern recognition. We explain the concrete frameworks and illustrate the potential superiority to prevent gradient explosion of our hybrid models. A group of numerical experiments about generalization power shows that our hybrid models possess the same generalization power as the pure classical models with different noisy datasets utilized. More importantly, another group of numerical experiments of robustness demonstrates that our hybrid models outperform pure classical models notably in resistance to parameter attacks with various asymmetric noises. Also, an ablation study indicate that the recognition accuracy of our hybrid models is 2\%-3\% higher than that of the quantum neural network without residual or dense connection. Eventually, we discuss the application scenarios of our hybrid models by analyzing their computational complexities.

comment: 12 pages for main paper with a hyperlink of a 18-page supplementary material in the last page of the main paper

♻ ☆ Analysis of Linear Mode Connectivity via Permutation-Based Weight Matching

Akira Ito, Masanori Yamada, Atsutoshi Kumagai

Recently, Ainsworth et al. showed that using weight matching (WM) to minimize the $L_2$ distance in a permutation search of model parameters effectively identifies permutations that satisfy linear mode connectivity (LMC), where the loss along a linear path between two independently trained models with different seeds remains nearly constant. This paper analyzes LMC using WM, which is useful for understanding stochastic gradient descent's effectiveness and its application in areas like model merging. We first empirically show that permutations found by WM do not significantly reduce the $L_2$ distance between two models, and the occurrence of LMC is not merely due to distance reduction by WM itself. We then demonstrate that permutations can change the directions of the singular vectors, but not the singular values, of the weight matrices in each layer. This finding shows that permutations found by WM primarily align the directions of singular vectors associated with large singular values across models. This alignment brings the singular vectors with large singular values, which determine the model's functionality, closer between the original and merged models, allowing the merged model to retain functionality similar to the original models, thereby satisfying LMC. This paper also analyzes activation matching (AM) in terms of singular vectors and finds that the principle of AM is the same as that of WM. Finally, we analyze the difference between WM and the straight-through estimator (STE), a dataset-dependent permutation search method, and show that WM can be more advantageous than STE in achieving LMC among three or more models.

comment: 36 pages

♻ ☆ Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation EMNLP 2024

Jirui Qi, Gabriele Sarti, Raquel Fernández, Arianna Bisazza

comment: Accepted by EMNLP 2024 Main Conference. Code and data released at https://github.com/Betswish/MIRAGE

Multimedia 8

☆ BadCM: Invisible Backdoor Attack Against Cross-Modal Learning

Zheng Zhang, Xu Yuan, Lei Zhu, Jingkuan Song, Liqiang Nie

☆ CalliffusionV2: Personalized Natural Calligraphy Generation with Flexible Multi-modal Control

Qisheng Liao, Liang Li, Yulang Fei, Gus Xia

In this paper, we introduce CalliffusionV2, a novel system designed to produce natural Chinese calligraphy with flexible multi-modal control. Unlike previous approaches that rely solely on image or text inputs and lack fine-grained control, our system leverages both images to guide generations at fine-grained levels and natural language texts to describe the features of generations. CalliffusionV2 excels at creating a broad range of characters and can quickly learn new styles through a few-shot learning approach. It is also capable of generating non-Chinese characters without prior training. Comprehensive tests confirm that our system produces calligraphy that is both stylistically accurate and recognizable by neural network classifiers and human evaluators.

comment: 11 pages, 7 figures

♻ ☆ SonicSense: Object Perception from In-Hand Acoustic Vibration

Jiaxun Liu, Boyuan Chen

We introduce SonicSense, a holistic design of hardware and software to enable rich robot object perception through in-hand acoustic vibration sensing. While previous studies have shown promising results with acoustic sensing for object perception, current solutions are constrained to a handful of objects with simple geometries and homogeneous materials, single-finger sensing, and mixing training and testing on the same objects. SonicSense enables container inventory status differentiation, heterogeneous material prediction, 3D shape reconstruction, and object re-identification from a diverse set of 83 real-world objects. Our system employs a simple but effective heuristic exploration policy to interact with the objects as well as end-to-end learning-based algorithms to fuse vibration signals to infer object properties. Our framework underscores the significance of in-hand acoustic vibration sensing in advancing robot tactile perception.

comment: Our project website is at: http://generalroboticslab.com/SonicSense

♻ ☆ Releasing the Parameter Latency of Neural Representation for High-Efficiency Video Compression

Gai Zhang, Xinfeng Zhang, Lv Tang, Yue Li, Kai Zhang, Li Zhang

♻ ☆ LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via a Hybrid Architecture

Xidong Wang, Dingjie Song, Shunian Chen, Chen Zhang, Benyou Wang

comment: 20 pages, 9 figures, 9 tables

♻ ☆ Bootstrap3D: Improving Multi-view Diffusion Model with Synthetic Data

Zeyi Sun, Tong Wu, Pan Zhang, Yuhang Zang, Xiaoyi Dong, Yuanjun Xiong, Dahua Lin, Jiaqi Wang

comment: Project Page: https://sunzey.github.io/Bootstrap3D/

♻ ☆ Semantic-Aware Adversarial Training for Reliable Deep Hashing Retrieval

Xu Yuan, Zheng Zhang, Xunguang Wang, Lin Wu

♻ ☆ Integrating Large Language Models into a Tri-Modal Architecture for Automated Depression Classification

Santosh V. Patapati

comment: Keywords: Multi-Modal Neural Networks, Deep Learning, Large Language Models, Depression Diagnosis, Biomedical Informatics, DAIC-WOZ

Performance 1

♻ ☆ GPUDrive: Data-driven, multi-agent driving simulation at 1 million FPS

Saman Kazemkhani, Aarav Pandya, Daphne Cornelisse, Brennan Shacklett, Eugene Vinitsky

2024-10-02

Hardware Architecture 5

☆ KeyVisor -- A Lightweight ISA Extension for Protected Key Handles with CPU-enforced Usage Policies

Fabian Schwarz, Jan Philipp Thoma, Christian Rossow, Tim Güneysu

The confidentiality of cryptographic keys is essential for the security of protection schemes used for communication, file encryption, and outsourced computation. Beyond cryptanalytic attacks, adversaries can steal keys from memory via software exploits or side channels, enabling them to, e.g., tamper with secrets or impersonate key owners. Therefore, existing defenses protect keys in dedicated devices or isolated memory, or store them only in encrypted form. However, these designs often provide unfavorable tradeoffs, sacrificing performance, fine-grained access control, or deployability. In this paper, we present KeyVisor, a lightweight ISA extension that securely offloads the handling of cryptographic keys to the CPU. KeyVisor provides CPU instructions that enable applications to request protected key handles and perform AEAD cipher operations on them. The underlying keys are accessible only by KeyVisor, and thus never leak to memory. KeyVisor's direct CPU integration enables fast crypto operations and hardware-enforced key usage restrictions, e.g., keys usable only for de-/encryption, with a limited lifetime, or with a process binding. Furthermore, privileged software, e.g., the monitor firmware of TEEs, can revoke keys or bind them to a specific process/TEE. We implement KeyVisor for RISC-V based on Rocket Chip, evaluate its performance, and demonstrate real-world use cases, including key-value databases, automotive feature licensing, and a read-only network middlebox.

comment: preprint

☆ Using a Performance Model to Implement a Superscalar CVA6

Côme Allart, Jean-Roch Coulon, André Sintzoff, Olivier Potin, Jean-Baptiste Rigaud

A performance model of CVA6 RISC-V processor is built to evaluate performance related modifications before implementing them in RTL. Its accuracy is 99.2% on CoreMark. This model is used to evaluate a superscalar feature for CVA6. During design phase, the model helped detecting and fixing performance bugs. The superscalar feature resulted in a CVA6 performance improvement of 40% on CoreMark.

♻ ☆ Do Not Trust Power Management: A Survey on Internal Energy-based Attacks Circumventing Trusted Execution Environments Security Properties

Owen Le Gonidec, Maria Méndez Real, Guillaume Bouffard, Jean-Christophe Prévotet

♻ ☆ ShortCircuit: AlphaZero-Driven Circuit Design

Dimitrios Tsaras, Antoine Grosnit, Lei Chen, Zhiyao Xie, Haitham Bou-Ammar, Mingxuan Yuan

Chip design relies heavily on generating Boolean circuits, such as AND-Inverter Graphs (AIGs), from functional descriptions like truth tables. This generation operation is a key process in logic synthesis, a primary chip design stage. While recent advances in deep learning have aimed to accelerate circuit design, these efforts have mostly focused on tasks other than synthesis, and traditional heuristic methods have plateaued. In this paper, we introduce ShortCircuit, a novel transformer-based architecture that leverages the structural properties of AIGs and performs efficient space exploration. Contrary to prior approaches attempting end-to-end generation of logic circuits using deep networks, ShortCircuit employs a two-phase process combining supervised with reinforcement learning to enhance generalization to unseen truth tables. We also propose an AlphaZero variant to handle the double exponentially large state space and the reward sparsity, enabling the discovery of near-optimal designs. To evaluate the generative performance of our model , we extract 500 truth tables from a set of 20 real-world circuits. ShortCircuit successfully generates AIGs for $98\%$ of the 8-input test truth tables, and outperforms the state-of-the-art logic synthesis tool, ABC, by $18.62\%$ in terms of circuits size.

♻ ☆ Design and In-training Optimization of Binary Search ADC for Flexible Classifiers SP

Paula Carolina Lozano Duarte, Florentia Afentaki, Georgios Zervakis, Mehdi B. Tahoori

Flexible Electronics (FE) offer distinct advantages, including mechanical flexibility and low process temperatures, enabling extremely low-cost production. To address the demands of applications such as smart sensors and wearables, flexible devices must be small and operate at low supply voltages. Additionally, target applications often require classifiers to operate directly on analog sensory input, necessitating the use of Analog to Digital Converters (ADCs) to process the sensory data. However, ADCs present serious challenges, particularly in terms of high area and power consumption, especially when considering stringent area and energy budget. In this work, we target common classifiers in this domain such as MLPs and SVMs and present a holistic approach to mitigate the elevated overhead of analog to digital interfacing in FE. First, we propose a novel design for Binary Search ADC that reduces area overhead 2X compared with the state-of-the-art Binary design and up to 5.4X compared with Flash ADC. Next, we present an in-training ADC optimization in which we keep the bare-minimum representations required and simplifying ADCs by removing unnecessary components. Our in-training optimization further reduces on average the area in terms of transistor count of the required ADCs by 5X for less than 1% accuracy loss.

comment: Accepted for publication at the 30th Asia and South Pacific Design Automation Conference (ASPDAC '25). doi: https://doi.org/10.1145/3658617.3697715

Distributed, Parallel, and Cluster Computing 24

☆ XChainWatcher: Monitoring and Identifying Attacks in Cross-Chain Bridges

André Augusto, Rafael Belchior, Jonas Pfannschmidt, André Vasconcelos, Miguel Correia

Cross-chain bridges are widely used blockchain interoperability mechanisms. However, several of these bridges have vulnerabilities that have caused 3.2 billion dollars in losses since May 2021. Some studies have revealed the existence of these vulnerabilities, but little quantitative research is available, and there are no safeguard mechanisms to protect bridges from such attacks. We propose XChainWatcher, the first mechanism for monitoring bridges and detecting attacks against them. XChainWatcher relies on a cross-chain model powered by a Datalog engine, designed to be pluggable into any cross-chain bridge. Analyzing data from the Ronin and Nomad bridges, we successfully identified the transactions that led to losses of \$611M and \$190M USD, respectively. XChainWatcher not only uncovers successful attacks but also reveals unintended behavior, such as 37 cross-chain transactions (cctx) that these bridges should not have accepted, failed attempts to exploit Nomad, over \$7.8M locked on one chain but never released on Ethereum, and \$200K lost due to inadequate interaction with bridges. We provide the first open-source dataset of 81,000 cctxs across three blockchains, capturing \$585M and \$3.7B in token transfers in Nomad and Ronin, respectively.

comment: 20 pages, 13 figures

☆ On the Resilience of Fast Failover Routing Against Dynamic Link Failures

Wenkai Dai, Klaus-Tycho Foerster, Stefan Schmid

Modern communication networks feature local fast failover mechanisms in the data plane, to swiftly respond to link failures with pre-installed rerouting rules. This paper explores resilient routing meant to tolerate $\leq k$ simultaneous link failures, ensuring packet delivery, provided that the source and destination remain connected. While past theoretical works studied failover routing under static link failures, i.e., links which permanently and simultaneously fail, real-world networks often face link flapping--dynamic down states caused by, e.g., numerous short-lived software-related faults. Thus, in this initial work, we re-investigate the resilience of failover routing against link flapping, by categorizing link failures into static, semi-dynamic (removing the assumption that links fail simultaneously), and dynamic (removing the assumption that links fail permanently) types, shedding light on the capabilities and limitations of failover routing under these scenarios. We show that $k$-edge-connected graphs exhibit $(k-1)$-resilient routing against dynamic failures for $k \leq 5$. We further show that this result extends to arbitrary $k$ if it is possible to rewrite $\log k$ bits in the packet header. Rewriting $3$ bits suffices to cope with $k$ semi-dynamic failures. However, on general graphs, tolerating $2$ dynamic failures becomes impossible without bit-rewriting. Even by rewriting $\log k$ bits, resilient routing cannot resolve $k$ dynamic failures, demonstrating the limitation of local fast rerouting.

☆ Decentralized Collaborative Inertial Tracking

Alpha Diallo, Benoit Garbinato

Although people spend most of their time indoors, outdoor tracking systems, such as the Global Positioning System (GPS), are predominantly used for location-based services. These systems are accurate outdoors, easy to use, and operate autonomously on each mobile device. In contrast, Indoor Tracking Systems~(ITS) lack standardization and are often difficult to operate because they require costly infrastructure. In this paper, we propose an indoor tracking algorithm that uses collected data from inertial sensors embedded in most mobile devices. In this setting, mobile devices autonomously estimate their location, hence removing the burden of deploying and maintaining complex and scattered hardware infrastructure. In addition, these devices collaborate by anonymously exchanging data with other nearby devices, using wireless communication, such as Bluetooth, to correct errors in their location estimates. Our collaborative algorithm relies on low-complexity geometry operations and can be deployed on any recent mobile device with commercial-grade sensors. We evaluate our solution on real-life data collected by different devices. Experimentation with 16 simultaneously moving and collaborating devices shows an average accuracy improvement of 44% compared to the standalone Pedestrian Dead Reckoning algorithm.

comment: ACCEPTED FOR PUBLICATION AND PRESENTED IN EAI MOBIQUITOUS 2023

☆ Generic Multicast(Extended Version)

José Augusto Bolina, Pierre Sutra, Douglas Antunes Rocha, Lasaro Camargos

comment: 12 pages, 2 figures

☆ Constant pH Simulation with FMM Electrostatics in GROMACS. (B) GPU Accelerated Hamiltonian Interpolation

Bartosz Kohnke, Eliane Briand, Carsten Kutzner, Helmut Grubmüller

The structural dynamics of biological macromolecules, such as proteins, DNA/RNA, or their complexes, are strongly influenced by protonation changes of their typically many titratable groups, which explains their pH sensitivity. In turn, conformational and environmental changes in the biomolecule affect the protonation state of these groups. With a few exceptions, conventional force field-based molecular dynamics (MD) simulations do not account for these effects, nor do they allow for coupling to a pH buffer. The $\lambda$-dynamics method implements this coupling and thus allows for MD simulations at constant pH. It uses separate Hamiltonians for the protonated and deprotonated states of each titratable group, with a $\lambda$ variable that continuously interpolates between them. However, rigorous implementations of Hamiltonian Interpolation (HI) $\lambda$-dynamics are prohibitively slow when used with Particle Mesh Ewald (PME). To circumvent this problem, it has been proposed to interpolate the charges instead of the Hamiltonians (QI). Here, we propose a rigorous yet efficient Multipole-Accelerated Hamiltonian Interpolation (MAHI) method to perform $\lambda$-dynamics in GROMACS. Starting from a charge-scaled Hamiltonian, precomputed with the Fast Multipole Method (FMM) or with PME, the correct HI forces are calculated with negligible computational overhead. We compare HI with QI and show that HI leads to more frequent transitions between protonation states, resulting in better sampling and accuracy. Our performance benchmarks show that introducing, e.g., 512 titratable sites to a one million atom MD system increases runtime by less than 20% compared to a regular FMM-based simulation. We have integrated the scheme into our GPU-FMM code for the simulation software GROMACS, allowing an easy and effortless transition from standard force field simulations to constant pH simulations.

☆ Scalable and Consistent Graph Neural Networks for Distributed Mesh-based Data-driven Modeling

Shivam Barwey, Riccardo Balin, Bethany Lusch, Saumil Patel, Ramesh Balakrishnan, Pinaki Pal, Romit Maulik, Venkatram Vishwanath

This work develops a distributed graph neural network (GNN) methodology for mesh-based modeling applications using a consistent neural message passing layer. As the name implies, the focus is on enabling scalable operations that satisfy physical consistency via halo nodes at sub-graph boundaries. Here, consistency refers to the fact that a GNN trained and evaluated on one rank (one large graph) is arithmetically equivalent to evaluations on multiple ranks (a partitioned graph). This concept is demonstrated by interfacing GNNs with NekRS, a GPU-capable exascale CFD solver developed at Argonne National Laboratory. It is shown how the NekRS mesh partitioning can be linked to the distributed GNN training and inference routines, resulting in a scalable mesh-based data-driven modeling workflow. We study the impact of consistency on the scalability of mesh-based GNNs, demonstrating efficient scaling in consistent GNNs for up to O(1B) graph nodes on the Frontier exascale supercomputer.

☆ A Novel Framework of Horizontal-Vertical Hybrid Federated Learning for EdgeIoT

Kai Li, Yilei Liang, Xin Yuan, Wei Ni, Jon Crowcroft, Chau Yuen, Ozgur B. Akan

This letter puts forth a new hybrid horizontal-vertical federated learning (HoVeFL) for mobile edge computing-enabled Internet of Things (EdgeIoT). In this framework, certain EdgeIoT devices train local models using the same data samples but analyze disparate data features, while the others focus on the same features using non-independent and identically distributed (non-IID) data samples. Thus, even though the data features are consistent, the data samples vary across devices. The proposed HoVeFL formulates the training of local and global models to minimize the global loss function. Performance evaluations on CIFAR-10 and SVHN datasets reveal that the testing loss of HoVeFL with 12 horizontal FL devices and six vertical FL devices is 5.5% and 25.2% higher, respectively, compared to a setup with six horizontal FL devices and 12 vertical FL devices.

comment: 5 pages, 3 figures

☆ Constant pH Simulation with FMM Electrostatics in GROMACS. (A) Design and Applications

Eliane Briand, Bartosz Kohnke, Carsten Kutzner, Helmut Grubmüller

The structural dynamics of biological macromolecules, such as proteins, DNA/RNA, or complexes thereof, are strongly influenced by protonation changes of their typically many titratable groups, which explains their sensitivity to pH changes. Conversely, conformational and environmental changes of the biomolecule affect the protonation state of these groups. With few exceptions, conventional force field-based molecular dynamics (MD) simulations do not account for these effects, nor do they allow for coupling to a pH buffer. Here we present a GROMACS implementation of a rigorous Hamiltonian interpolation $\lambda$-dynamics constant pH method, which rests on GPU-accelerated Fast Multipole Method (FMM) electrostatics. Our implementation supports both CHARMM36m and Amber99sb*-ILDN force fields and is largely automated to enable seamless switching from regular MD to constant pH MD, involving minimal changes to the input files. Here, the first of two companion papers describes the underlying constant pH protocol and sample applications to several prototypical benchmark systems such as cardiotoxin V, lysozyme, and staphylococcal nuclease. Enhanced convergence is achieved through a new dynamic barrier height optimization method, and high p$K_a$ accuracy is demonstrated. We use Functional Mode Analysis and Mutual Information to explore the complex intra- and intermolecular couplings between the protonation states of titratable groups as well as those between protonation states and conformational dynamics. We identify striking conformation-dependent p$K_a$ variations and unexpected inter-residue couplings. Conformation-protonation coupling is identified as a primary cause of the slow protonation convergence notorious to constant pH simulations involving multiple titratable groups, suggesting enhanced sampling methods to accelerate convergence.

☆ Multi-level Memory-Centric Profiling on ARM Processors with ARM SPE SC

Samuel Miksits, Ruimin Shi, Maya Gokhale, Jacob Wahlgren, Gabin Schieffer, Ivy Peng

High-end ARM processors are emerging in data centers and HPC systems, posing as a strong contender to x86 machines. Memory-centric profiling is an important approach for dissecting an application's bottlenecks on memory access and guiding optimizations. Many existing memory profiling tools leverage hardware performance counters and precise event sampling, such as Intel PEBS and AMD IBS, to achieve high accuracy and low overhead. In this work, we present a multi-level memory profiling tool for ARM processors, leveraging Statistical Profiling Extension (SPE). We evaluate the tool using both HPC and Cloud workloads on the ARM Ampere processor. Our results provide the first quantitative assessment of time overhead and sampling accuracy of ARM SPE for memory-centric profiling at different sampling periods and aux buffer sizes.

comment: To be published in Workshop Proceedings of The International Conference for High Performance Computing Networking, Storage, and Analysis (SC-W '24) (2024)

☆ Personalized Federated Learning on Flowing Data Heterogeneity under Restricted Storage

Sixing Tan, Xianmin Liu

Recent years, researchers focused on personalized federated learning (pFL) to address the inconsistent requirements of clients causing by data heterogeneity in federated learning (FL). However, existing pFL methods typically assume that local data distribution remains unchanged during FL training, the changing data distribution in actual heterogeneous data scenarios can affect model convergence rate and reduce model performance. In this paper, we focus on solving the pFL problem under the situation where data flows through each client like a flowing stream which called Flowing Data Heterogeneity under Restricted Storage, and shift the training goal to the comprehensive performance of the model throughout the FL training process. Therefore, based on the idea of category decoupling, we design a local data distribution reconstruction scheme and a related generator architecture to reduce the error of the controllable replayed data distribution, then propose our pFL framework, pFedGRP, to achieve knowledge transfer and personalized aggregation. Comprehensive experiments on five datasets with multiple settings show the superiority of pFedGRP over eight baseline methods.

☆ Theoretical Lower Bounds for the Oven Scheduling Problem

Francesca Da Ros, Marie-Louise Lackner, Nysret Musliu

The Oven Scheduling Problem (OSP) is an NP-hard real-world parallel batch scheduling problem arising in the semiconductor industry. The objective of the problem is to schedule a set of jobs on ovens while minimizing several factors, namely total oven runtime, job tardiness, and setup costs. At the same time, it must adhere to various constraints such as oven eligibility and availability, job release dates, setup times between batches, and oven capacity limitations. The key to obtaining efficient schedules is to process compatible jobs simultaneously in batches. In this paper, we develop theoretical, problem-specific lower bounds for the OSP that can be computed very quickly. We thoroughly examine these lower bounds, evaluating their quality and exploring their integration into existing solution methods. Specifically, we investigate their contribution to exact methods and a metaheuristic local search approach using simulated annealing. Moreover, these problem-specific lower bounds enable us to assess the solution quality for large instances for which exact methods often fail to provide tight lower bounds.

comment: arXiv admin note: text overlap with arXiv:2203.12517

☆ ParallelSFL: A Novel Split Federated Learning Framework Tackling Heterogeneity Issues

Yunming Liao, Yang Xu, Hongli Xu, Zhiwei Yao, Liusheng Huang, Chunming Qiao

Mobile devices contribute more than half of the world's web traffic, providing massive and diverse data for powering various federated learning (FL) applications. In order to avoid the communication bottleneck on the parameter server (PS) and accelerate the training of large-scale models on resourceconstraint workers in edge computing (EC) system, we propose a novel split federated learning (SFL) framework, termed ParallelSFL. Concretely, we split an entire model into a bottom submodel and a top submodel, and divide participating workers into multiple clusters, each of which collaboratively performs the SFL training procedure and exchanges entire models with the PS. However, considering the statistical and system heterogeneity in edge systems, it is challenging to arrange suitable workers to specific clusters for efficient model training. To address these challenges, we carefully develop an effective clustering strategy by optimizing a utility function related to training efficiency and model accuracy. Specifically, ParallelSFL partitions workers into different clusters under the heterogeneity restrictions, thereby promoting model accuracy as well as training efficiency. Meanwhile, ParallelSFL assigns diverse and appropriate local updating frequencies for each cluster to further address system heterogeneity. Extensive experiments are conducted on a physical platform with 80 NVIDIA Jetson devices, and the experimental results show that ParallelSFL can reduce the traffic consumption by at least 21%, speed up the model training by at least 1.36x, and improve model accuracy by at least 5% in heterogeneous scenarios, compared to the baselines.

comment: arXiv admin note: text overlap with arXiv:2311.13348

☆ ConServe: Harvesting GPUs for Low-Latency and High-Throughput Large Language Model Serving

Yifan Qiao, Shu Anzai, Shan Yu, Haoran Ma, Yang Wang, Miryung Kim, Harry Xu

Many applications are leveraging large language models (LLMs) for complex tasks, and they generally demand low inference latency and high serving throughput for interactive online jobs such as chatbots. However, the tight latency requirement and high load variance of applications pose challenges to serving systems in achieving high GPU utilization. Due to the high costs of scheduling and preemption, today's systems generally use separate clusters to serve online and offline inference tasks, and dedicate GPUs for online inferences to avoid interference. This approach leads to underutilized GPUs because one must reserve enough GPU resources for the peak expected load, even if the average load is low. This paper proposes to harvest stranded GPU resources for offline LLM inference tasks such as document summarization and LLM benchmarking. Unlike online inferences, these tasks usually run in a batch-processing manner with loose latency requirements, making them a good fit for stranded resources that are only available shortly. To enable safe and efficient GPU harvesting without interfering with online tasks, we built ConServe, an LLM serving system that contains (1) an execution engine that preempts running offline tasks upon the arrival of online tasks, (2) an incremental checkpointing mechanism that minimizes the amount of recomputation required by preemptions, and (3) a scheduler that adaptively batches offline tasks for higher GPU utilization. Our evaluation demonstrates that ConServe achieves strong performance isolation when co-serving online and offline tasks but at a much higher GPU utilization. When colocating practical online and offline workloads on popular models such as Llama-2-7B, ConServe achieves 2.35$\times$ higher throughput than state-of-the-art online serving systems and reduces serving latency by 84$\times$ compared to existing co-serving systems.

☆ Exploring Fine-grained Task Parallelism on Simultaneous Multithreading Cores

Denis Los, Igor Petushkov

Nowadays, latency-critical, high-performance applications are parallelized even on power-constrained client systems to improve performance. However, an important scenario of fine-grained tasking on simultaneous multithreading CPU cores in such systems has not been well researched in previous works. Hence, in this paper, we conduct performance analysis of state-of-the-art shared-memory parallel programming frameworks on simultaneous multithreading cores using real-world fine-grained application kernels. We introduce a specialized and simple software-only parallel programming framework called Relic to enable extremely fine-grained tasking on simultaneous multithreading cores. Using Relic framework, we increase performance speedups over serial implementations of benchmark kernels by 19.1% compared to LLVM OpenMP, by 31.0% compared to GNU OpenMP, by 20.2% compared to Intel OpenMP, by 33.2% compared to X-OpenMP, by 30.1% compared to oneTBB, by 23.0% compared to Taskflow, and by 21.4% compared to OpenCilk.

☆ Debiasing Federated Learning with Correlated Client Participation

Zhenyu Sun, Ziyang Zhang, Zheng Xu, Gauri Joshi, Pranay Sharma, Ermin Wei

In cross-device federated learning (FL) with millions of mobile clients, only a small subset of clients participate in training in every communication round, and Federated Averaging (FedAvg) is the most popular algorithm in practice. Existing analyses of FedAvg usually assume the participating clients are independently sampled in each round from a uniform distribution, which does not reflect real-world scenarios. This paper introduces a theoretical framework that models client participation in FL as a Markov chain to study optimization convergence when clients have non-uniform and correlated participation across rounds. We apply this framework to analyze a more general and practical pattern: every client must wait a minimum number of $R$ rounds (minimum separation) before re-participating. We theoretically prove and empirically observe that increasing minimum separation reduces the bias induced by intrinsic non-uniformity of client availability in cross-device FL systems. Furthermore, we develop an effective debiasing algorithm for FedAvg that provably converges to the unbiased optimal solution under arbitrary minimum separation and unknown client availability distribution.

☆ Learning the Optimal Path and DNN Partition for Collaborative Edge Inference

Yin Huang, Letian Zhang, Jie Xu

Recent advancements in Deep Neural Networks (DNNs) have catalyzed the development of numerous intelligent mobile applications and services. However, they also introduce significant computational challenges for resource-constrained mobile devices. To address this, collaborative edge inference has been proposed. This method involves partitioning a DNN inference task into several subtasks and distributing these across multiple network nodes. Despite its potential, most current approaches presume known network parameters -- like node processing speeds and link transmission rates -- or rely on a fixed sequence of nodes for processing the DNN subtasks. In this paper, we tackle a more complex scenario where network parameters are unknown and must be learned, and multiple network paths are available for distributing inference tasks. Specifically, we explore the learning problem of selecting the optimal network path and assigning DNN layers to nodes along this path, considering potential security threats and the costs of switching paths. We begin by deriving structural insights from the DNN layer assignment with complete network information, which narrows down the decision space and provides crucial understanding of optimal assignments. We then cast the learning problem with incomplete network information as a novel adversarial group linear bandits problem with switching costs, featuring rewards generation through a combined stochastic and adversarial process. We introduce a new bandit algorithm, B-EXPUCB, which combines elements of the classical blocked EXP3 and LinUCB algorithms, and demonstrate its sublinear regret. Extensive simulations confirm B-EXPUCB's superior performance in learning for collaborative edge inference over existing algorithms.

comment: 15 pages, 15 figures, submitted to IEEE journals for possible publication

☆ Real-World Data and Calibrated Simulation Suite for Offline Training of Reinforcement Learning Agents to Optimize Energy and Emission in Buildings for Environmental Sustainability

Judah Goldfeder, John Sipple

Commercial office buildings contribute 17 percent of Carbon Emissions in the US, according to the US Energy Information Administration (EIA), and improving their efficiency will reduce their environmental burden and operating cost. A major contributor of energy consumption in these buildings are the Heating, Ventilation, and Air Conditioning (HVAC) devices. HVAC devices form a complex and interconnected thermodynamic system with the building and outside weather conditions, and current setpoint control policies are not fully optimized for minimizing energy use and carbon emission. Given a suitable training environment, a Reinforcement Learning (RL) agent is able to improve upon these policies, but training such a model, especially in a way that scales to thousands of buildings, presents many practical challenges. Most existing work on applying RL to this important task either makes use of proprietary data, or focuses on expensive and proprietary simulations that may not be grounded in the real world. We present the Smart Buildings Control Suite, the first open source interactive HVAC control dataset extracted from live sensor measurements of devices in real office buildings. The dataset consists of two components: six years of real-world historical data from three buildings, for offline RL, and a lightweight interactive simulator for each of these buildings, calibrated using the historical data, for online and model-based RL. For ease of use, our RL environments are all compatible with the OpenAI gym environment standard. We also demonstrate a novel method of calibrating the simulator, as well as baseline results on training an RL agent on the simulator, predicting real-world data, and training an RL agent directly from data. We believe this benchmark will accelerate progress and collaboration on building optimization and environmental sustainability research.

♻ ☆ Advancing Hybrid Defense for Byzantine Attacks in Federated Learning

Kai Yue, Richeng Jin, Chau-Wai Wong, Huaiyu Dai

Federated learning (FL) enables multiple clients to collaboratively train a global model without sharing their local data. Recent studies have highlighted the vulnerability of FL to Byzantine attacks, where malicious clients send poisoned updates to degrade model performance. Notably, many attacks have been developed targeting specific aggregation rules, whereas various defense mechanisms have been designed for dedicated threat models. This paper studies the resilience of an attack-agnostic FL scenario, where the server lacks prior knowledge of both the attackers' strategies and the number of malicious clients involved. We first introduce a hybrid defense against state-of-the-art attacks. Our goal is to identify a general-purpose aggregation rule that performs well on average while also avoiding worst-case vulnerabilities. By adaptively selecting from available defenses, we demonstrate that the server remains robust even when confronted with a substantial proportion of poisoned updates. To better understand this resilience, we then assess the attackers' capability using a proxy called client heterogeneity. We also emphasize that the existing FL defenses should not be regarded as secure, as demonstrated through the newly proposed Trapsetter attack. The proposed attack outperforms other state-of-the-art attacks by further reducing the model test accuracy by 8-10%. Our findings highlight the ongoing need for the development of Byzantine-resilient aggregation algorithms in FL.

♻ ☆ Asynchronous Federated Reinforcement Learning with Policy Gradient Updates: Algorithm Design and Convergence Analysis

Guangchen Lan, Dong-Jun Han, Abolfazl Hashemi, Vaneet Aggarwal, Christopher G. Brinton

To improve the efficiency of reinforcement learning (RL), we propose a novel asynchronous federated reinforcement learning (FedRL) framework termed AFedPG, which constructs a global model through collaboration among $N$ agents using policy gradient (PG) updates. To address the challenge of lagged policies in asynchronous settings, we design a delay-adaptive lookahead technique \textit{specifically for FedRL} that can effectively handle heterogeneous arrival times of policy gradients. We analyze the theoretical global convergence bound of AFedPG, and characterize the advantage of the proposed algorithm in terms of both the sample complexity and time complexity. Specifically, our AFedPG method achieves $O(\frac{{\epsilon}^{-2.5}}{N})$ sample complexity for global convergence at each agent on average. Compared to the single agent setting with $O(\epsilon^{-2.5})$ sample complexity, it enjoys a linear speedup with respect to the number of agents. Moreover, compared to synchronous FedPG, AFedPG improves the time complexity from $O(\frac{t_{\max}}{N})$ to $O({\sum_{i=1}^{N} \frac{1}{t_{i}}})^{-1}$, where $t_{i}$ denotes the time consumption in each iteration at agent $i$, and $t_{\max}$ is the largest one. The latter complexity $O({\sum_{i=1}^{N} \frac{1}{t_{i}}})^{-1}$ is always smaller than the former one, and this improvement becomes significant in large-scale federated settings with heterogeneous computing powers ($t_{\max}\gg t_{\min}$). Finally, we empirically verify the improved performance of AFedPG in four widely-used MuJoCo environments with varying numbers of agents. We also demonstrate the advantages of AFedPG in various computing heterogeneity scenarios.

♻ ☆ TACOS: Topology-Aware Collective Algorithm Synthesizer for Distributed Machine Learning

William Won, Midhilesh Elavazhagan, Sudarshan Srinivasan, Swati Gupta, Tushar Krishna

The surge of artificial intelligence, particularly large language models, has driven the rapid development of large-scale machine learning clusters. Executing distributed models on these clusters is often constrained by communication overhead, making efficient utilization of available network resources crucial. As a result, the routing algorithm employed for collective communications (i.e., collective algorithms) plays a pivotal role in determining overall performance. Unfortunately, existing collective communication libraries for distributed machine learning are limited by a fixed set of basic collective algorithms. This limitation hinders communication optimization, especially in modern clusters with heterogeneous and asymmetric topologies. Furthermore, manually designing collective algorithms for all possible combinations of network topologies and collective patterns requires heavy engineering and validation efforts. To address these challenges, this paper presents TACOS, an autonomous synthesizer capable of automatically generating topology-aware collective algorithms tailored to specific collective patterns and network topologies. TACOS is highly flexible, synthesizing an All-Reduce algorithm for a heterogeneous 128-NPU system in just 1.08 seconds, while achieving up to a 4.27x performance improvement over state-of-the-art synthesizers. Additionally, TACOS demonstrates better scalability with polynomial synthesis times, in contrast to NP-hard approaches which only scale to systems with tens of NPUs. TACOS can synthesize for 40K NPUs in just 2.52 hours.

comment: Contains 12 main pages, 21 figures, 5 tables. Artifact appendix attached

♻ ☆ Leveraging MTD to Mitigate Poisoning Attacks in Decentralized FL with Non-IID Data

Chao Feng, Alberto Huertas Celdrán, Zien Zeng, Zi Ye, Jan von der Assen, Gerome Bovet, Burkhard Stiller

Decentralized Federated Learning (DFL), a paradigm for managing big data in a privacy-preserved manner, is still vulnerable to poisoning attacks where malicious clients tamper with data or models. Current defense methods often assume Independently and Identically Distributed (IID) data, which is unrealistic in real-world applications. In non-IID contexts, existing defensive strategies face challenges in distinguishing between models that have been compromised and those that have been trained on heterogeneous data distributions, leading to diminished efficacy. In response, this paper proposes a framework that employs the Moving Target Defense (MTD) approach to bolster the robustness of DFL models. By continuously modifying the attack surface of the DFL system, this framework aims to mitigate poisoning attacks effectively. The proposed MTD framework includes both proactive and reactive modes, utilizing a reputation system that combines metrics of model similarity and loss, alongside various defensive techniques. Comprehensive experimental evaluations indicate that the MTD-based mechanism significantly mitigates a range of poisoning attack types across multiple datasets with different topologies.

♻ ☆ Federated Instruction Tuning of LLMs with Domain Coverage Augmentation

Zezhou Wang, Yaxin Du, Zhuzhong Qian, Siheng Chen

Federated Domain-specific Instruction Tuning (FedDIT) utilizes limited cross-client private data together with server-side public data for instruction augmentation, ultimately boosting model performance within specific domains. To date, the factors affecting FedDIT remain unclear, and existing instruction augmentation methods primarily focus on the centralized setting without considering distributed environments. Our experiments reveal that the cross-client domain coverage, rather than data heterogeneity, drives model performance in FedDIT. In response, we propose FedDCA, which optimizes domain coverage through greedy client center selection and retrieval-based augmentation. For client-side computational efficiency and system scalability, FedDCA$^*$, the variant of FedDCA, utilizes heterogeneous encoders with server-side feature alignment. Extensive experiments across four distinct domains (code, medical, financial, and mathematical) substantiate the effectiveness of both methods. Additionally, we investigate privacy preservation against memory extraction attacks utilizing various amounts of public data. Results show that there is no significant correlation between the volume of public data and the privacy-preserving capability. However, as the fine-tuning rounds increase, the risk of privacy leakage reduces or converges.

♻ ☆ HybridFlow: A Flexible and Efficient RLHF Framework

Guangming Sheng, Chi Zhang, Zilingfeng Ye, Xibin Wu, Wang Zhang, Ru Zhang, Yanghua Peng, Haibin Lin, Chuan Wu

Reinforcement Learning from Human Feedback (RLHF) is widely used in Large Language Model (LLM) alignment. Traditional RL can be modeled as a dataflow, where each node represents computation of a neural network (NN) and each edge denotes data dependencies between the NNs. RLHF complicates the dataflow by expanding each node into a distributed LLM training or generation program, and each edge into a many-to-many multicast. Traditional RL frameworks execute the dataflow using a single controller to instruct both intra-node computation and inter-node communication, which can be inefficient in RLHF due to large control dispatch overhead for distributed intra-node computation. Existing RLHF systems adopt a multi-controller paradigm, which can be inflexible due to nesting distributed computation and data communication. We propose HybridFlow, which combines single-controller and multi-controller paradigms in a hybrid manner to enable flexible representation and efficient execution of the RLHF dataflow. We carefully design a set of hierarchical APIs that decouple and encapsulate computation and data dependencies in the complex RLHF dataflow, allowing efficient operation orchestration to implement RLHF algorithms and flexible mapping of the computation onto various devices. We further design a 3D-HybridEngine for efficient actor model resharding between training and generation phases, with zero memory redundancy and significantly reduced communication overhead. Our experimental results demonstrate 1.53$\times$~20.57$\times$ throughput improvement when running various RLHF algorithms using HybridFlow, as compared with state-of-the-art baselines. HybridFlow source code will be available at https://github.com/volcengine/verl.

♻ ☆ FPGA-based Distributed Union-Find Decoder for Surface Codes

Namitha Liyanage, Yue Wu, Siona Tagare, Lin Zhong

A fault-tolerant quantum computer must decode and correct errors faster than they appear to prevent exponential slowdown due to error correction. The Union-Find (UF) decoder is promising with an average time complexity slightly higher than $O(d^3)$. We report a distributed version of the UF decoder that exploits parallel computing resources for further speedup. Using an FPGA-based implementation, we empirically show that this distributed UF decoder has a sublinear average time complexity with regard to $d$, given $O(d^3)$ parallel computing resources. The decoding time per measurement round decreases as $d$ increases, the first time for a quantum error decoder. The implementation employs a scalable architecture called Helios that organizes parallel computing resources into a hybrid tree-grid structure. Using a Xilinx VCU129 FPGA, we successfully implement $d$ up to 21 with an average decoding time of 11.5 ns per measurement round under 0.1\% phenomenological noise, and 23.7 ns for $d=17$ under equivalent circuit-level noise. This performance is significantly faster than any existing decoder implementation. Furthermore, we show that Helios can optimize for resource efficiency by decoding $d=51$ on a Xilinx VCU129 FPGA with an average latency of 544ns per measurement round.

comment: The article extends the work in arXiv:2301.08419, which also appeared in https://ieeexplore.ieee.org/document/10313800

Operating Systems 2

☆ Global Scheduling of Weakly-Hard Real-Time Tasks using Job-Level Priority Classes

V. Gabriel Moyano, Zain A. H. Hammadeh, Selma Saidi, Daniel Lüdtke

Real-time systems are intrinsic components of many pivotal applications, such as self-driving vehicles, aerospace and defense systems. The trend in these applications is to incorporate multiple tasks onto fewer, more powerful hardware platforms, e.g., multi-core systems, mainly for reducing cost and power consumption. Many real-time tasks, like control tasks, can tolerate occasional deadline misses due to robust algorithms. These tasks can be modeled using the weakly-hard model. Literature shows that leveraging the weakly-hard model can relax the over-provisioning associated with designed real-time systems. However, a wide-range of the research focuses on single-core platforms. Therefore, we strive to extend the state-of-the-art of scheduling weakly-hard real-time tasks to multi-core platforms. We present a global job-level fixed priority scheduling algorithm together with its schedulability analysis. The scheduling algorithm leverages the tolerable continuous deadline misses to assigning priorities to jobs. The proposed analysis extends the Response Time Analysis (RTA) for global scheduling to test the schedulability of tasks. Hence, our analysis scales with the number of tasks and number of cores because, unlike literature, it depends neither on Integer Linear Programming nor reachability trees. Schedulability analyses show that the schedulability ratio is improved by 40% comparing to the global Rate Monotonic (RM) scheduling and up to 60% more than the global EDF scheduling, which are the state-of-the-art schedulers on the RTEMS real-time operating system. Our evaluation on industrial embedded multi-core platform running RTEMS shows that the scheduling overhead of our proposal does not exceed 60 Nanosecond.

♻ ☆ CyberCortex.AI: An AI-based Operating System for Autonomous Robotics and Complex Automation

Sorin Grigorescu, Mihai Zaha

The underlying framework for controlling autonomous robots and complex automation applications are Operating Systems (OS) capable of scheduling perception-and-control tasks, as well as providing real-time data communication to other robotic peers and remote cloud computers. In this paper, we introduce CyberCortex AI, a robotics OS designed to enable heterogeneous AI-based robotics and complex automation applications. CyberCortex AI is a decentralized distributed OS which enables robots to talk to each other, as well as to High Performance Computers (HPC) in the cloud. Sensory and control data from the robots is streamed towards HPC systems with the purpose of training AI algorithms, which are afterwards deployed on the robots. Each functionality of a robot (e.g. sensory data acquisition, path planning, motion control, etc.) is executed within a so-called DataBlock of Filters shared through the internet, where each filter is computed either locally on the robot itself, or remotely on a different robotic system. The data is stored and accessed via a so-called Temporal Addressable Memory (TAM), which acts as a gateway between each filter's input and output. CyberCortex.AI has two main components: i) the CyberCortex AI inference system, which is a real-time implementation of the DataBlock running on the robots' embedded hardware, and ii) the CyberCortex AI dojo, which runs on an HPC computer in the cloud, and it is used to design, train and deploy AI algorithms. We present a quantitative and qualitative performance analysis of the proposed approach using two collaborative robotics applications: i) a forest fires prevention system based on an Unitree A1 legged robot and an Anafi Parrot 4K drone, as well as ii) an autonomous driving system which uses CyberCortex.AI for collaborative perception and motion control.

Programming Languages 5

☆ Surveying the Rust Verification Landscape

Alex Le Blanc, Patrick Lam

Rust aims to be a safe programming language applicable to systems programming applications. In particular, its type system has strong guardrails to prevent a variety of issues, such as memory safety bugs and data races. However, these guardrails can be sidestepped via the unsafe keyword. unsafe allows certain otherwise-prohibited operations, but shifts the onus of preventing undefined behaviour from the Rust language's compile-time checks to the developer. We believe that tools have a role to play in ensuring the absence of undefined behaviour in the presence of unsafe code. Moreover, safety aside, programs would also benefit from being verified for functional correctness, ensuring that they meet their specifications. In this research proposal, we explore what it means to do Rust verification. Specifically, we explore which properties are worth verifying for Rust; what techniques exist to verify them; and which code is worth verifying. In doing so, we motivate an effort to verify safety properties of the Rust standard library, presenting the relevant challenges along with ideas to address them.

☆ SecCoder: Towards Generalizable and Robust Secure Code Generation EMNLP

Boyu Zhang, Tianyu Du, Junkai Tong, Xuhong Zhang, Kingsum Chow, Sheng Cheng, Xun Wang, Jianwei Yin

After large models (LMs) have gained widespread acceptance in code-related tasks, their superior generative capacity has greatly promoted the application of the code LM. Nevertheless, the security of the generated code has raised attention to its potential damage. Existing secure code generation methods have limited generalizability to unseen test cases and poor robustness against the attacked model, leading to safety failures in code generation. In this paper, we propose a generalizable and robust secure code generation method SecCoder by using in-context learning (ICL) and the safe demonstration. The dense retriever is also used to select the most helpful demonstration to maximize the improvement of the generated code's security. Experimental results show the superior generalizability of the proposed model SecCoder compared to the current secure code generation method, achieving a significant security improvement of an average of 7.20% on unseen test cases. The results also show the better robustness of SecCoder compared to the current attacked code LM, achieving a significant security improvement of an average of 7.74%. Our analysis indicates that SecCoder enhances the security of LMs in generating code, and it is more generalizable and robust.

comment: To Appear in the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP)

☆ From Code to Correctness: Closing the Last Mile of Code Generation with Hierarchical Debugging

Yuling Shi, Songsong Wang, Chengcheng Wan, Xiaodong Gu

comment: Code and data available at https://github.com/YerbaPage/MGDebugger

♻ ☆ ScenicNL: Generating Probabilistic Scenario Programs from Natural Language

Karim Elmaaroufi, Devan Shanker, Ana Cismaru, Marcell Vazquez-Chanlatte, Alberto Sangiovanni-Vincentelli, Matei Zaharia, Sanjit A. Seshia

For cyber-physical systems (CPS), including robotics and autonomous vehicles, mass deployment has been hindered by fatal errors that occur when operating in rare events. To replicate rare events such as vehicle crashes, many companies have created logging systems and employed crash reconstruction experts to meticulously recreate these valuable events in simulation. However, in these methods, "what if" questions are not easily formulated and answered. We present ScenarioNL, an AI System for creating scenario programs from natural language. Specifically, we generate these programs from police crash reports. Reports normally contain uncertainty about the exact details of the incidents which we represent through a Probabilistic Programming Language (PPL), Scenic. By using Scenic, we can clearly and concisely represent uncertainty and variation over CPS behaviors, properties, and interactions. We demonstrate how commonplace prompting techniques with the best Large Language Models (LLM) are incapable of reasoning about probabilistic scenario programs and generating code for low-resource languages such as Scenic. Our system is comprised of several LLMs chained together with several kinds of prompting strategies, a compiler, and a simulator. We evaluate our system on publicly available autonomous vehicle crash reports in California from the last five years and share insights into how we generate code that is both semantically meaningful and syntactically correct.

comment: 22 pages, 3 figures. Published at COLM 2024. https://ke7.github.io/ScenicNL

♻ ☆ WatChat: Explaining perplexing programs by debugging mental models ACL

Kartik Chandra, Katherine M. Collins, Will Crichton, Tony Chen, Tzu-Mao Li, Adrian Weller, Rachit Nigam, Joshua Tenenbaum, Jonathan Ragan-Kelley

Often, a good explanation for a program's unexpected behavior is a bug in the programmer's code. But sometimes, an even better explanation is a bug in the programmer's mental model of the language or API they are using. Instead of merely debugging our current code ("giving the programmer a fish"), what if our tools could directly debug our mental models ("teaching the programmer to fish")? In this paper, we apply recent ideas from computational cognitive science to offer a principled framework for doing exactly that. Given a "why?" question about a program, we automatically infer potential misconceptions about the language/API that might cause the user to be surprised by the program's behavior -- and then analyze those misconceptions to provide explanations of the program's behavior. Our key idea is to formally represent misconceptions as counterfactual (erroneous) semantics for the language/API, which can be inferred and debugged using program synthesis techniques. We demonstrate our framework, WatChat, by building systems for explanation in two domains: JavaScript type coercion, and the Git version control system. We evaluate WatChatJS and WatChatGit by comparing their outputs to experimentally-collected human-written explanations in these two domains: we show that WatChat's explanations exhibit key features of human-written explanation, unlike those of a state-of-the-art language model.

comment: This is a preprint of work presented in early-stage non-archival form at the ACL Natural Language Reasoning and Structured Explanations Workshop

Computer Vision and Pattern Recognition 19

☆ MVGS: Multi-view-regulated Gaussian Splatting for Novel View Synthesis

Xiaobiao Du, Yida Wang, Xin Yu

Recent works in volume rendering, \textit{e.g.} NeRF and 3D Gaussian Splatting (3DGS), significantly advance the rendering quality and efficiency with the help of the learned implicit neural radiance field or 3D Gaussians. Rendering on top of an explicit representation, the vanilla 3DGS and its variants deliver real-time efficiency by optimizing the parametric model with single-view supervision per iteration during training which is adopted from NeRF. Consequently, certain views are overfitted, leading to unsatisfying appearance in novel-view synthesis and imprecise 3D geometries. To solve aforementioned problems, we propose a new 3DGS optimization method embodying four key novel contributions: 1) We transform the conventional single-view training paradigm into a multi-view training strategy. With our proposed multi-view regulation, 3D Gaussian attributes are further optimized without overfitting certain training views. As a general solution, we improve the overall accuracy in a variety of scenarios and different Gaussian variants. 2) Inspired by the benefit introduced by additional views, we further propose a cross-intrinsic guidance scheme, leading to a coarse-to-fine training procedure concerning different resolutions. 3) Built on top of our multi-view regulated training, we further propose a cross-ray densification strategy, densifying more Gaussian kernels in the ray-intersect regions from a selection of views. 4) By further investigating the densification strategy, we found that the effect of densification should be enhanced when certain views are distinct dramatically. As a solution, we propose a novel multi-view augmented densification strategy, where 3D Gaussians are encouraged to get densified to a sufficient number accordingly, resulting in improved reconstruction accuracy.

comment: Project Page:https://xiaobiaodu.github.io/mvgs-project/

☆ Orient Anything

Christopher Scarvelis, David Benhaim, Paul Zhang

Orientation estimation is a fundamental task in 3D shape analysis which consists of estimating a shape's orientation axes: its side-, up-, and front-axes. Using this data, one can rotate a shape into canonical orientation, where its orientation axes are aligned with the coordinate axes. Developing an orientation algorithm that reliably estimates complete orientations of general shapes remains an open problem. We introduce a two-stage orientation pipeline that achieves state of the art performance on up-axis estimation and further demonstrate its efficacy on full-orientation estimation, where one seeks all three orientation axes. Unlike previous work, we train and evaluate our method on all of Shapenet rather than a subset of classes. We motivate our engineering contributions by theory describing fundamental obstacles to orientation estimation for rotationally-symmetric shapes, and show how our method avoids these obstacles.

☆ EC-DIT: Scaling Diffusion Transformers with Adaptive Expert-Choice Routing

Haotian Sun, Bowen Zhang, Yanghao Li, Haoshuo Huang, Tao Lei, Ruoming Pang, Bo Dai, Nan Du

Diffusion transformers have been widely adopted for text-to-image synthesis. While scaling these models up to billions of parameters shows promise, the effectiveness of scaling beyond current sizes remains underexplored and challenging. By explicitly exploiting the computational heterogeneity of image generations, we develop a new family of Mixture-of-Experts (MoE) models (EC-DIT) for diffusion transformers with expert-choice routing. EC-DIT learns to adaptively optimize the compute allocated to understand the input texts and generate the respective image patches, enabling heterogeneous computation aligned with varying text-image complexities. This heterogeneity provides an efficient way of scaling EC-DIT up to 97 billion parameters and achieving significant improvements in training convergence, text-to-image alignment, and overall generation quality over dense models and conventional MoE models. Through extensive ablations, we show that EC-DIT demonstrates superior scalability and adaptive compute allocation by recognizing varying textual importance through end-to-end training. Notably, in text-to-image alignment evaluation, our largest models achieve a state-of-the-art GenEval score of 71.68% and still maintain competitive inference speed with intuitive interpretability.

☆ Tracking objects that change in appearance with phase synchrony

Sabine Muzellec, Drew Linsley, Alekh K. Ashok, Ennio Mingolla, Girik Malik, Rufin VanRullen, Thomas Serre

Objects we encounter often change appearance as we interact with them. Changes in illumination (shadows), object pose, or movement of nonrigid objects can drastically alter available image features. How do biological visual systems track objects as they change? It may involve specific attentional mechanisms for reasoning about the locations of objects independently of their appearances -- a capability that prominent neuroscientific theories have associated with computing through neural synchrony. We computationally test the hypothesis that the implementation of visual attention through neural synchrony underlies the ability of biological visual systems to track objects that change in appearance over time. We first introduce a novel deep learning circuit that can learn to precisely control attention to features separately from their location in the world through neural synchrony: the complex-valued recurrent neural network (CV-RNN). Next, we compare object tracking in humans, the CV-RNN, and other deep neural networks (DNNs), using FeatureTracker: a large-scale challenge that asks observers to track objects as their locations and appearances change in precisely controlled ways. While humans effortlessly solved FeatureTracker, state-of-the-art DNNs did not. In contrast, our CV-RNN behaved similarly to humans on the challenge, providing a computational proof-of-concept for the role of phase synchronization as a neural substrate for tracking appearance-morphing objects as they move about.

☆ Anchors Aweigh! Sail for Optimal Unified Multi-Modal Representations

Minoh Jeong, Min Namgung, Zae Myung Kim, Dongyeop Kang, Yao-Yi Chiang, Alfred Hero

Multimodal learning plays a crucial role in enabling machine learning models to fuse and utilize diverse data sources, such as text, images, and audio, to support a variety of downstream tasks. A unified representation across various modalities is particularly important for improving efficiency and performance. Recent binding methods, such as ImageBind (Girdhar et al., 2023), typically use a fixed anchor modality to align multimodal data in the anchor modal embedding space. In this paper, we mathematically analyze the fixed anchor binding methods and uncover notable limitations: (1) over-reliance on the choice of the anchor modality, (2) failure to capture intra-modal information, and (3) failure to account for inter-modal correlation among non-anchored modalities. To address these limitations, we propose CentroBind, a simple yet powerful approach that eliminates the need for a fixed anchor; instead, it employs dynamically adjustable centroid-based anchors generated from all available modalities, resulting in a balanced and rich representation space. We theoretically demonstrate that our method captures three crucial properties of multimodal learning: intra-modal learning, inter-modal learning, and multimodal alignment, while also constructing a robust unified representation across all modalities. Our experiments on both synthetic and real-world datasets demonstrate the superiority of the proposed method, showing that dynamic anchor methods outperform all fixed anchor binding methods as the former captures more nuanced multimodal interactions.

☆ EMMA: Efficient Visual Alignment in Multi-Modal LLMs

Sara Ghazanfari, Alexandre Araujo, Prashanth Krishnamurthy, Siddharth Garg, Farshad Khorrami

Multi-modal Large Language Models (MLLMs) have recently exhibited impressive general-purpose capabilities by leveraging vision foundation models to encode the core concepts of images into representations. These are then combined with instructions and processed by the language model to generate high-quality responses. Despite significant progress in enhancing the language component, challenges persist in optimally fusing visual encodings within the language model for task-specific adaptability. Recent research has focused on improving this fusion through modality adaptation modules but at the cost of significantly increased model complexity and training data needs. In this paper, we propose EMMA (Efficient Multi-Modal Adaptation), a lightweight cross-modality module designed to efficiently fuse visual and textual encodings, generating instruction-aware visual representations for the language model. Our key contributions include: (1) an efficient early fusion mechanism that integrates vision and language representations with minimal added parameters (less than 0.2% increase in model size), (2) an in-depth interpretability analysis that sheds light on the internal mechanisms of the proposed method; (3) comprehensive experiments that demonstrate notable improvements on both specialized and general benchmarks for MLLMs. Empirical results show that EMMA boosts performance across multiple tasks by up to 9.3% while significantly improving robustness against hallucinations. Our code is available at https://github.com/SaraGhazanfari/EMMA

☆ Posterior sampling via Langevin dynamics based on generative priors

Vishal Purohit, Matthew Repasky, Jianfeng Lu, Qiang Qiu, Yao Xie, Xiuyuan Cheng

Posterior sampling in high-dimensional spaces using generative models holds significant promise for various applications, including but not limited to inverse problems and guided generation tasks. Despite many recent developments, generating diverse posterior samples remains a challenge, as existing methods require restarting the entire generative process for each new sample, making the procedure computationally expensive. In this work, we propose efficient posterior sampling by simulating Langevin dynamics in the noise space of a pre-trained generative model. By exploiting the mapping between the noise and data spaces which can be provided by distilled flows or consistency models, our method enables seamless exploration of the posterior without the need to re-run the full sampling chain, drastically reducing computational overhead. Theoretically, we prove a guarantee for the proposed noise-space Langevin dynamics to approximate the posterior, assuming that the generative model sufficiently approximates the prior distribution. Our framework is experimentally validated on image restoration tasks involving noisy linear and nonlinear forward operators applied to LSUN-Bedroom (256 x 256) and ImageNet (64 x 64) datasets. The results demonstrate that our approach generates high-fidelity samples with enhanced semantic diversity even under a limited number of function evaluations, offering superior efficiency and performance compared to existing diffusion-based posterior sampling techniques.

☆ Kolmogorov-Arnold Network Autoencoders

Mohammadamin Moradi, Shirin Panahi, Erik Bollt, Ying-Cheng Lai

Deep learning models have revolutionized various domains, with Multi-Layer Perceptrons (MLPs) being a cornerstone for tasks like data regression and image classification. However, a recent study has introduced Kolmogorov-Arnold Networks (KANs) as promising alternatives to MLPs, leveraging activation functions placed on edges rather than nodes. This structural shift aligns KANs closely with the Kolmogorov-Arnold representation theorem, potentially enhancing both model accuracy and interpretability. In this study, we explore the efficacy of KANs in the context of data representation via autoencoders, comparing their performance with traditional Convolutional Neural Networks (CNNs) on the MNIST, SVHN, and CIFAR-10 datasets. Our results demonstrate that KAN-based autoencoders achieve competitive performance in terms of reconstruction accuracy, thereby suggesting their viability as effective tools in data analysis tasks.

comment: 12 pages, 5 figures, 1 table

☆ Depth Pro: Sharp Monocular Metric Depth in Less Than a Second

Aleksei Bochkovskii, Amaël Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R. Richter, Vladlen Koltun

We present a foundation model for zero-shot metric monocular depth estimation. Our model, Depth Pro, synthesizes high-resolution depth maps with unparalleled sharpness and high-frequency details. The predictions are metric, with absolute scale, without relying on the availability of metadata such as camera intrinsics. And the model is fast, producing a 2.25-megapixel depth map in 0.3 seconds on a standard GPU. These characteristics are enabled by a number of technical contributions, including an efficient multi-scale vision transformer for dense prediction, a training protocol that combines real and synthetic datasets to achieve high metric accuracy alongside fine boundary tracing, dedicated evaluation metrics for boundary accuracy in estimated depth maps, and state-of-the-art focal length estimation from a single image. Extensive experiments analyze specific design choices and demonstrate that Depth Pro outperforms prior work along multiple dimensions. We release code and weights at https://github.com/apple/ml-depth-pro

comment: Code and weights available at https://github.com/apple/ml-depth-pro

☆ Learning from the Giants: A Practical Approach to Underwater Depth and Surface Normals Estimation

Alzayat Saleh, Melanie Olsen, Bouchra Senadji, Mostafa Rahimi Azghadi

Monocular Depth and Surface Normals Estimation (MDSNE) is crucial for tasks such as 3D reconstruction, autonomous navigation, and underwater exploration. Current methods rely either on discriminative models, which struggle with transparent or reflective surfaces, or generative models, which, while accurate, are computationally expensive. This paper presents a novel deep learning model for MDSNE, specifically tailored for underwater environments, using a hybrid architecture that integrates Convolutional Neural Networks (CNNs) with Transformers, leveraging the strengths of both approaches. Training effective MDSNE models is often hampered by noisy real-world datasets and the limited generalization of synthetic datasets. To address this, we generate pseudo-labeled real data using multiple pre-trained MDSNE models. To ensure the quality of this data, we propose the Depth Normal Evaluation and Selection Algorithm (DNESA), which evaluates and selects the most reliable pseudo-labeled samples using domain-specific metrics. A lightweight student model is then trained on this curated dataset. Our model reduces parameters by 90% and training costs by 80%, allowing real-time 3D perception on resource-constrained devices. Key contributions include: a novel and efficient MDSNE model, the DNESA algorithm, a domain-specific data pipeline, and a focus on real-time performance and scalability. Designed for real-world underwater applications, our model facilitates low-cost deployments in underwater robots and autonomous vehicles, bridging the gap between research and practical implementation.

comment: 18 pages, 6 figures, 8 tables. Submitted to Elsevier

☆ Semi-Supervised Fine-Tuning of Vision Foundation Models with Content-Style Decomposition

Mariia Drozdova, Vitaliy Kinakh, Yury Belousov, Erica Lastufka, Slava Voloshynovskiy

In this paper, we present a semi-supervised fine-tuning approach designed to improve the performance of foundation models on downstream tasks with limited labeled data. By leveraging content-style decomposition within an information-theoretic framework, our method enhances the latent representations of pre-trained vision foundation models, aligning them more effectively with specific task objectives and addressing the problem of distribution shift. We evaluate our approach on multiple datasets, including MNIST, its augmented variations (with yellow and white stripes), CIFAR-10, SVHN, and GalaxyMNIST. The experiments show improvements over purely supervised baselines, particularly in low-labeled data regimes, across both frozen and trainable backbones for the majority of the tested datasets.

☆ DisEnvisioner: Disentangled and Enriched Visual Prompt for Customized Image Generation

Jing He, Haodong Li, Yongzhe Hu, Guibao Shen, Yingjie Cai, Weichao Qiu, Ying-Cong Chen

In the realm of image generation, creating customized images from visual prompt with additional textual instruction emerges as a promising endeavor. However, existing methods, both tuning-based and tuning-free, struggle with interpreting the subject-essential attributes from the visual prompt. This leads to subject-irrelevant attributes infiltrating the generation process, ultimately compromising the personalization quality in both editability and ID preservation. In this paper, we present DisEnvisioner, a novel approach for effectively extracting and enriching the subject-essential features while filtering out -irrelevant information, enabling exceptional customization performance, in a tuning-free manner and using only a single image. Specifically, the feature of the subject and other irrelevant components are effectively separated into distinctive visual tokens, enabling a much more accurate customization. Aiming to further improving the ID consistency, we enrich the disentangled features, sculpting them into more granular representations. Experiments demonstrate the superiority of our approach over existing methods in instruction response (editability), ID consistency, inference speed, and the overall image quality, highlighting the effectiveness and efficiency of DisEnvisioner. Project page: https://disenvisioner.github.io/.

comment: The first two authors contributed equally. Project page: https://disenvisioner.github.io/

☆ Using Style Ambiguity Loss to Improve Aesthetics of Diffusion Models

James Baker

Teaching text-to-image models to be creative involves using style ambiguity loss. In this work, we explore using the style ambiguity training objective, used to approximate creativity, on a diffusion model. We then experiment with forms of style ambiguity loss that do not require training a classifier or a labeled dataset, and find that the models trained with style ambiguity loss can generate better images than the baseline diffusion models and GANs. Code is available at https://github.com/jamesBaker361/clipcreate.

comment: arXiv admin note: substantial text overlap with arXiv:2407.12009

☆ Improving Autonomous AI Agents with Reflective Tree Search and Self-Learning

Xiao Yu, Baolin Peng, Vineeth Vajipey, Hao Cheng, Michel Galley, Jianfeng Gao, Zhou Yu

Autonomous agents have demonstrated significant potential in automating complex multistep decision-making tasks. However, even state-of-the-art vision-language models (VLMs), such as GPT-4o, still fall short of human-level performance, particularly in intricate web environments and long-horizon planning tasks. To address these limitations, we introduce Reflective Monte Carlo Tree Search (R-MCTS), a novel test-time algorithm designed to enhance the ability of AI agents, e.g., powered by GPT-4o, to explore decision space on the fly. R-MCTS extends traditional MCTS by 1) incorporating contrastive reflection, allowing agents to learn from past interactions and dynamically improve their search efficiency; and 2) using multi-agent debate to provide reliable state evaluation. Moreover, we improve the agent's performance by fine-tuning GPT-4o through self-learning, using R-MCTS generated tree traversals without any human-provided labels. On the challenging VisualWebArena benchmark, our GPT-4o-based R-MCTS agent achieves a 6% to 30% relative improvement across various tasks compared to the previous state-of-the-art. Additionally, we show that the knowledge gained from test-time search can be effectively transferred back to GPT-4o via fine-tuning. The fine-tuned GPT-4o matches 97% of R-MCTS's performance while reducing compute usage by a factor of four at test time. Furthermore, qualitative results reveal that the fine-tuned GPT-4o model demonstrates the ability to explore the environment, evaluate a state, and backtrack to viable ones when it detects that the current state cannot lead to success. Moreover, our work demonstrates the compute scaling properties in both training - data collection with R-MCTS - and testing time. These results suggest a promising research direction to enhance VLMs' reasoning and planning capabilities for agentic applications via test-time search and self-learning.

♻ ☆ Generative Visual Instruction Tuning

Jefferson Hernandez, Ruben Villegas, Vicente Ordonez

We propose to use automatically generated instruction-following data to improve the zero-shot capabilities of a large multimodal model with additional support for generative and image editing tasks. We achieve this by curating a new multimodal instruction-following set using GPT-4V and existing datasets for image generation and editing. Using this instruction set and the existing LLaVA-Finetune instruction set for visual understanding tasks, we produce GenLLaVA, a Generative Large Language and Visual Assistant. GenLLaVA is built through a strategy that combines three types of large pretrained models through instruction finetuning: Mistral for language modeling, SigLIP for image-text matching, and StableDiffusion for text-to-image generation. Our model demonstrates visual understanding capabilities superior to LLaVA and additionally demonstrates competitive results with native multimodal models such as Unified-IO 2, paving the way for building advanced general-purpose visual assistants by effectively re-using existing multimodal models. We open-source our dataset, codebase, and model checkpoints to foster further research and application in this domain.

comment: Add more results using task tokens, expand the introduction and related work FIX: error in LLM-as-judge evaluation that was over-inflating the results

♻ ☆ Law of the Weakest Link: Cross Capabilities of Large Language Models

Ming Zhong, Aston Zhang, Xuewei Wang, Rui Hou, Wenhan Xiong, Chenguang Zhu, Zhengxing Chen, Liang Tan, Chloe Bi, Mike Lewis, Sravya Popuri, Sharan Narang, Melanie Kambadur, Dhruv Mahajan, Sergey Edunov, Jiawei Han, Laurens van der Maaten

The development and evaluation of Large Language Models (LLMs) have largely focused on individual capabilities. However, this overlooks the intersection of multiple abilities across different types of expertise that are often required for real-world tasks, which we term cross capabilities. To systematically explore this concept, we first define seven core individual capabilities and then pair them to form seven common cross capabilities, each supported by a manually constructed taxonomy. Building on these definitions, we introduce CrossEval, a benchmark comprising 1,400 human-annotated prompts, with 100 prompts for each individual and cross capability. To ensure reliable evaluation, we involve expert annotators to assess 4,200 model responses, gathering 8,400 human ratings with detailed explanations to serve as reference examples. Our findings reveal that, in both static evaluations and attempts to enhance specific abilities, current LLMs consistently exhibit the "Law of the Weakest Link," where cross-capability performance is significantly constrained by the weakest component. Specifically, across 58 cross-capability scores from 17 models, 38 scores are lower than all individual capabilities, while 20 fall between strong and weak, but closer to the weaker ability. These results highlight the under-performance of LLMs in cross-capability tasks, making the identification and improvement of the weakest capabilities a critical priority for future research to optimize performance in complex, multi-dimensional scenarios.

comment: Data, Code, & Benchmark: www.llm-cross-capabilities.org

♻ ☆ Lotus: Diffusion-based Visual Foundation Model for High-quality Dense Prediction

Jing He, Haodong Li, Wei Yin, Yixun Liang, Leheng Li, Kaiqiang Zhou, Hongbo Zhang, Bingbing Liu, Ying-Cong Chen

Leveraging the visual priors of pre-trained text-to-image diffusion models offers a promising solution to enhance zero-shot generalization in dense prediction tasks. However, existing methods often uncritically use the original diffusion formulation, which may not be optimal due to the fundamental differences between dense prediction and image generation. In this paper, we provide a systemic analysis of the diffusion formulation for the dense prediction, focusing on both quality and efficiency. And we find that the original parameterization type for image generation, which learns to predict noise, is harmful for dense prediction; the multi-step noising/denoising diffusion process is also unnecessary and challenging to optimize. Based on these insights, we introduce Lotus, a diffusion-based visual foundation model with a simple yet effective adaptation protocol for dense prediction. Specifically, Lotus is trained to directly predict annotations instead of noise, thereby avoiding harmful variance. We also reformulate the diffusion process into a single-step procedure, simplifying optimization and significantly boosting inference speed. Additionally, we introduce a novel tuning strategy called detail preserver, which achieves more accurate and fine-grained predictions. Without scaling up the training data or model capacity, Lotus achieves SoTA performance in zero-shot depth and normal estimation across various datasets. It also enhances efficiency, being significantly faster than most existing diffusion-based methods. Lotus' superior quality and efficiency also enable a wide range of practical applications, such as joint estimation, single/multi-view 3D reconstruction, etc. Project page: https://lotus3d.github.io/.

comment: The first two authors contributed equally. Project page: https://lotus3d.github.io/

♻ ☆ ViC-MAE: Self-Supervised Representation Learning from Images and Video with Contrastive Masked Autoencoders ECCV 2024

Jefferson Hernandez, Ruben Villegas, Vicente Ordonez

We propose ViC-MAE, a model that combines both Masked AutoEncoders (MAE) and contrastive learning. ViC-MAE is trained using a global featured obtained by pooling the local representations learned under an MAE reconstruction loss and leveraging this representation under a contrastive objective across images and video frames. We show that visual representations learned under ViC-MAE generalize well to both video and image classification tasks. Particularly, ViC-MAE obtains state-of-the-art transfer learning performance from video to images on Imagenet-1k compared to the recently proposed OmniMAE by achieving a top-1 accuracy of 86% (+1.3% absolute improvement) when trained on the same data and 87.1% (+2.4% absolute improvement) when training on extra data. At the same time ViC-MAE outperforms most other methods on video benchmarks by obtaining 75.9% top-1 accuracy on the challenging Something something-v2 video benchmark . When training on videos and images from a diverse combination of datasets, our method maintains a balanced transfer-learning performance between video and image classification benchmarks, coming only as a close second to the best supervised method.

comment: Published at ECCV 2024

♻ ☆ CLIP-MoE: Towards Building Mixture of Experts for CLIP with Diversified Multiplet Upcycling

Jihai Zhang, Xiaoye Qu, Tong Zhu, Yu Cheng

In recent years, Contrastive Language-Image Pre-training (CLIP) has become a cornerstone in multimodal intelligence. However, recent studies have identified that the information loss in the CLIP encoding process is substantial, and CLIP tends to capture only coarse-grained features from the input. This deficiency significantly limits the ability of a single CLIP model to handle images rich in visual detail. In this work, we propose a simple yet effective model-agnostic strategy, Diversified Multiplet Upcycling (DMU), for CLIP. DMU efficiently fine-tunes a series of CLIP models that capture different feature spaces, from a dense pre-trained CLIP checkpoint, sharing parameters except for the Feed-Forward Network (FFN). These models can then be transformed into a CLIP-MoE with a larger model capacity, leading to significantly enhanced performance with minimal computational overhead. To the best of our knowledge, Diversified Multiplet Upcycling is the first approach to introduce sparsely activated MoE into CLIP foundation models. Extensive experiments demonstrate the significant performance of CLIP-MoE across various zero-shot retrieval, zero-shot image classification tasks, and downstream Multimodal Large Language Model (MLLM) benchmarks by serving as a vision encoder. Furthermore, Diversified Multiplet Upcycling enables the conversion of any dense CLIP model into CLIP-MoEs, which can seamlessly replace CLIP in a plug-and-play manner without requiring further adaptation in downstream frameworks. Through Diversified Multiplet Upcycling, we aim to provide valuable insights for future research on developing more efficient and effective multimodal learning systems.

Multimedia 6

☆ Social Media Authentication and Combating Deepfakes using Semi-fragile Invisible Image Watermarking

Aakash Varma Nadimpalli, Ajita Rattani

With the significant advances in deep generative models for image and video synthesis, Deepfakes and manipulated media have raised severe societal concerns. Conventional machine learning classifiers for deepfake detection often fail to cope with evolving deepfake generation technology and are susceptible to adversarial attacks. Alternatively, invisible image watermarking is being researched as a proactive defense technique that allows media authentication by verifying an invisible secret message embedded in the image pixels. A handful of invisible image watermarking techniques introduced for media authentication have proven vulnerable to basic image processing operations and watermark removal attacks. In response, we have proposed a semi-fragile image watermarking technique that embeds an invisible secret message into real images for media authentication. Our proposed watermarking framework is designed to be fragile to facial manipulations or tampering while being robust to benign image-processing operations and watermark removal attacks. This is facilitated through a unique architecture of our proposed technique consisting of critic and adversarial networks that enforce high image quality and resiliency to watermark removal efforts, respectively, along with the backbone encoder-decoder and the discriminator networks. Thorough experimental investigations on SOTA facial Deepfake datasets demonstrate that our proposed model can embed a $64$-bit secret as an imperceptible image watermark that can be recovered with a high-bit recovery accuracy when benign image processing operations are applied while being non-recoverable when unseen Deepfake manipulations are applied. In addition, our proposed watermarking technique demonstrates high resilience to several white-box and black-box watermark removal attacks. Thus, obtaining state-of-the-art performance.

comment: ACM Transactions (Digital Threats: Research and Practice)

☆ RADAR: Robust Two-stage Modality-incomplete Industrial Anomaly Detection

Bingchen Miao, Wenqiao Zhang, Juncheng Li, Siliang Tang, Zhaocheng Li, Haochen Shi, Jun Xiao, Yueting Zhuang

Multimodal Industrial Anomaly Detection (MIAD), utilizing 3D point clouds and 2D RGB images to identify the abnormal region of products, plays a crucial role in industrial quality inspection. However, the conventional MIAD setting presupposes that all 2D and 3D modalities are paired, overlooking the fact that multimodal data collected from the real world is often imperfect due to missing modalities. Consequently, MIAD models that demonstrate robustness against modal-incomplete data are highly desirable in practice. To address this practical challenge, we introduce a first-of-its-kind study that comprehensively investigates Modality-Incomplete Industrial Anomaly Detection (MIIAD), to consider the imperfect learning environment in which the multimodal information may be incomplete. Not surprisingly, we discovered that most existing MIAD approaches are inadequate for addressing MIIAD challenges, leading to significant performance degradation on the MIIAD benchmark we developed. In this paper, we propose a novel two-stage Robust modAlity-imcomplete fusing and Detecting frAmewoRk, abbreviated as RADAR. Our bootstrapping philosophy is to enhance two stages in MIIAD, improving the robustness of the Multimodal Transformer: i) In feature fusion, we first explore learning modality-incomplete instruction, guiding the pre-trained Multimodal Transformer to robustly adapt to various modality-incomplete scenarios, and implement adaptive parameter learning based on a HyperNetwork; ii) In anomaly detection, we construct a real-pseudo hybrid module to highlight the distinctiveness of modality combinations, further enhancing the robustness of the MIIAD model. Our experimental results demonstrate that the proposed RADAR significantly surpasses conventional MIAD methods in terms of effectiveness and robustness on our newly created MIIAD dataset, underscoring its practical application value.

☆ Harnessing the Latent Diffusion Model for Training-Free Image Style Transfer

Kento Masui, Mayu Otani, Masahiro Nomura, Hideki Nakayama

Diffusion models have recently shown the ability to generate high-quality images. However, controlling its generation process still poses challenges. The image style transfer task is one of those challenges that transfers the visual attributes of a style image to another content image. Typical obstacle of this task is the requirement of additional training of a pre-trained model. We propose a training-free style transfer algorithm, Style Tracking Reverse Diffusion Process (STRDP) for a pretrained Latent Diffusion Model (LDM). Our algorithm employs Adaptive Instance Normalization (AdaIN) function in a distinct manner during the reverse diffusion process of an LDM while tracking the encoding history of the style image. This algorithm enables style transfer in the latent space of LDM for reduced computational cost, and provides compatibility for various LDM models. Through a series of experiments and a user study, we show that our method can quickly transfer the style of an image without additional training. The speed, compatibility, and training-free aspect of our algorithm facilitates agile experiments with combinations of styles and LDMs for extensive application.

♻ ☆ Progressive Frame Patching for FoV-based Point Cloud Video Streaming

Tongyu Zong, Yixiang Mao, Chen Li, Yong Liu, Yao Wang

Many XR applications require the delivery of volumetric video to users with six degrees of freedom (6-DoF) movements. Point Cloud has become a popular volumetric video format. A dense point cloud consumes much higher bandwidth than a 2D/360 degree video frame. User Field of View (FoV) is more dynamic with 6-DoF movement than 3-DoF movement. To save bandwidth, FoV-adaptive streaming predicts a user's FoV and only downloads point cloud data falling in the predicted FoV. However, it is vulnerable to FoV prediction errors, which can be significant when a long buffer is utilized for smoothed streaming. In this work, we propose a multi-round progressive refinement framework for point cloud video streaming. Instead of sequentially downloading point cloud frames, our solution simultaneously downloads/patches multiple frames falling into a sliding time-window, leveraging the inherent scalability of octree-based point-cloud coding. The optimal rate allocation among all tiles of active frames are solved analytically using the heterogeneous tile rate-quality functions calibrated by the predicted user FoV. Multi-frame downloading/patching simultaneously takes advantage of the streaming smoothness resulting from long buffer and the FoV prediction accuracy at short buffer length. We evaluate our streaming solution using simulations driven by real point cloud videos, real bandwidth traces, and 6-DoF FoV traces of real users. Our solution is robust against the bandwidth/FoV prediction errors, and can deliver high and smooth view quality in the face of bandwidth variations and dynamic user and point cloud movements.

comment: Transactions on Multimedia (under review)

♻ ☆ DiffSSD: A Diffusion-Based Dataset For Speech Forensics ICASSP

Kratika Bhagtani, Amit Kumar Singh Yadav, Paolo Bestagini, Edward J. Delp

Diffusion-based speech generators are ubiquitous. These methods can generate very high quality synthetic speech and several recent incidents report their malicious use. To counter such misuse, synthetic speech detectors have been developed. Many of these detectors are trained on datasets which do not include diffusion-based synthesizers. In this paper, we demonstrate that existing detectors trained on one such dataset, ASVspoof2019, do not perform well in detecting synthetic speech from recent diffusion-based synthesizers. We propose the Diffusion-Based Synthetic Speech Dataset (DiffSSD), a dataset consisting of about 200 hours of labeled speech, including synthetic speech generated by 8 diffusion-based open-source and 2 commercial generators. We also examine the performance of existing synthetic speech detectors on DiffSSD in both closed-set and open-set scenarios. The results highlight the importance of this dataset in detecting synthetic speech generated from recent open-source and commercial speech generators.

comment: Submitted to IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP) 2025

♻ ☆ Beyond Single-Audio: Advancing Multi-Audio Processing in Audio Large Language Models EMNLP24

Yiming Chen, Xianghu Yue, Xiaoxue Gao, Chen Zhang, Luis Fernando D'Haro, Robby T. Tan, Haizhou Li

Various audio-LLMs (ALLMs) have been explored recently for tackling different audio tasks simultaneously using a single, unified model. While existing evaluations of ALLMs primarily focus on single-audio tasks, real-world applications often involve processing multiple audio streams simultaneously. To bridge this gap, we propose the first multi-audio evaluation (MAE) benchmark that consists of 20 datasets from 11 multi-audio tasks encompassing both speech and sound scenarios. Comprehensive experiments on MAE demonstrate that the existing ALLMs, while being powerful in comprehending primary audio elements in individual audio inputs, struggling to handle multi-audio scenarios. To this end, we propose a novel multi-audio-LLM (MALLM) to capture audio context among multiple similar audios using discriminative learning on our proposed synthetic data. The results demonstrate that the proposed MALLM outperforms all baselines and achieves high data efficiency using synthetic data without requiring human annotations. The proposed MALLM opens the door for ALLMs towards multi-audio processing era and brings us closer to replicating human auditory capabilities in machines.

comment: EMNLP24 Findings